Acute respiratory infections risk prediction using machine learning among Ethiopian children Aged 6 Months to 2 Years

Acute respiratory infections risk prediction using machine learning among Ethiopian children Aged 6 Months to 2 Years | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Acute respiratory infections risk prediction using machine learning among Ethiopian children Aged 6 Months to 2 Years Ewunate Assaye Kassaw, Biruk Beletew Abate, Ashenafi Kibret Sendekie This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7425776/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Introduction : Acute respiratory infections (ARI) caused by various pathogens are the cause of millions of illnesses and deaths among children under five. The prevalence of ARI is higher in low- and middle-income countries. To this date, in low- and middle-income countries, the management of ARI in children under the age of two is mainly curative, not preventive. Thus, this study aimed to explore the capability of machine learning models to predict the forthcoming ARI from the general demographic health survey data by developing and deploying predictive machine learning models. Methods : The demographic health survey data was obtained from the USAID repository, the data was preprocessed, and the important features were identified. Then data class balancing was done using synthetic minority oversampling techniques. Then, logistic regression, support vector machine, k-nearest neighbor, decision tree, random forest, gradient boosting, and one-dimensional convolutional neural network models were developed. The K-fold cross-validation technique was used to train the model and obtain a stable model and representative performance metrics. The accuracy, the recall, the F1 score, the precision, and the AUC score results were calculated and used to select the best-performing model. Finally, the selected model was deployed on Streamlit as a web-based application and using the Python tkinter library for developing desktop applications. Results : A total of 2500 subjects’ data were obtained, out of which 503 subjects were having coughs, which is nearly one-fifth of the total data. Upon applying the synthetic minority oversampling technique (SMOTE), the overall data is increased to 3992, with each class having 1996 subjects’ data. At first, the data had 23 features, but after changing some features from categories to numbers and giving numerical values to ordered and yes/no features, there were 36 features in total. Following data class balancing and data preprocessing, seven models were trained and resulted in AUC scores of 0.842, 0.881, 0.860, 0.792, 0.918, 0.918, 0.918, 0.726, and 0.872, and recall scores of 0.745, 0.790, 0.914, 0.827, 0.862, 0.716, and 0.824 were obtained for LR, SVM, KNN, DT, RF, GBC, and 1DCNN models, respectively. Then the best-performing model, which is the random forest model, was selected and deployed as a web-based application on Streamlit and as an offline Windows application using the Python tkinter library. Conclusion : This study illustrates the possibilities of machine learning backend applications for predicting the forthcoming ARI from the demographic health survey data, which will play a key role in preventing diseases upon necessary regulatory and quality checks. In low-resource setting areas that are highly vulnerable to ARI, machine learning-based applications will be useful. Further studies need to be done considering a wider range of parameters for improving the predictability and accuracy of the models. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction Acute respiratory infections (ARIs) are a group of diseases caused by a wide range of pathogens, including Streptococcus pneumoniae, Haemophilus influenzae, and Staphylococcus aureus.[ 1 ]. ARIs are a leading cause of illness and death in children worldwide[ 2 ], accounting for approximately 12 million cases and 1.3 million deaths annually among those under five years old[ 3 ]. ARIs are particularly prevalent in low- and middle-income countries (LMICs), where they remain among the top causes of childhood morbidity and mortality[ 4 ], [ 5 ], [ 6 ]. A significant proportion of these deaths occur in Sub-Saharan Africa and other developing regions[ 7 ], [ 8 ]. Contributing factors in LMICs include limited access to healthcare, malnutrition, poor sanitation, and inadequate immunization coverage [ 9 ]. Ethiopia is among the top 15 countries with the highest burden of acute respiratory infections (ARIs), affecting approximately four million children each year[ 10 ]. ARIs cause over 40,000 deaths annually among children under five, accounting for 18% of all child deaths in the country[ 11 ]. Pneumonia, a major form of ARI, contributes significantly to this mortality, driven by factors such as poverty, environmental pollution, and limited healthcare infrastructure[ 12 ]. Despite efforts to reduce mortality through vaccination programs and improved healthcare access, ARIs remain a major public health challenge in Ethiopia and similar settings[ 13 ]. Early and accurate prediction of ARIs is crucial for timely intervention and improved health outcomes[ 14 ]. Traditional diagnostic approaches rely on clinical assessment and laboratory tests, which may be time-consuming, costly, or inaccessible in resource-limited settings [ 15 ]. In recent years, machine learning (ML) has emerged as a promising tool for predicting and diagnosing various health conditions, including respiratory diseases, by leveraging patient data and identifying patterns that may not be apparent through conventional analysis[ 16 ], [ 17 ]. Previous studies have explored the application of ML in respiratory disease prediction, mainly focusing on older children and adults[ 18 ], [ 19 ], [ 20 ], [ 21 ]. Recent studies have also predicted the risk factors of ARIs in children using ML approaches[ 18 ], [ 19 ], [ 22 ], [ 23 ], [ 24 ], [ 25 ]. The most employed ML algorithms include Random Forest (RF), Extreme Gradient Boosting (XGBoost), Gradient Boosting, Naïve Bayes (NB), Logistic Regression (LR), and Decision Trees, and among these, ensemble models—particularly those combining RF [ 18 ], [ 19 ], [ 22 ] and XGBoost [ 25 ] have consistently demonstrated superior performance. Across the literature, several recurring predictors of ARIs have been identified. At the child level, key factors include duration of breastfeeding, history of diarrhea, immunization status, and nutritional status[ 18 ], [ 22 ], [ 23 ]. Maternal and household variables such as educational attainment, occupation, wealth status, residential setting (urban or rural), cooking fuel, and the number of living children have also been associated with ARI risk[ 18 ], [ 23 ]. Despite these advancements, the existing literature lacks age-specific modelling tailored to infants and young children aged 6 to 24 months, who represent a particularly vulnerable subgroup due to their developmental stage and immature immune systems. Most existing studies have focused broadly on children under five, thus overlooking the unique risk profiles of this younger under two years. The rationale for this study stems from the need to develop an early prediction system that can aid healthcare providers in identifying at-risk children before severe complications arise. Given the challenges of early clinical diagnosis in infants and toddlers, who may present with non-specific symptoms, ML can be a valuable decision-support tool. ML models can offer high predictive accuracy by utilizing a combination of clinical, demographic, and environmental factors, assisting in timely clinical decision-making and reducing unnecessary hospital admissions[ 26 ], [ 27 ]. Moreover, integrating such predictive models into community healthcare settings could enhance surveillance efforts and improve resource allocation in pediatric care[ 28 ]. Unbalanced use of demographic health survey data, inability to consider the difference in the regional location of the subjects in the country, considering only infected subjects, inability to predict early, and inability to prepare the developed model into a utility tool were the research gaps of the existing studies done on the prediction of ARI using machine learning for the Ethiopian population. Additionally, researchers have not yet studied ARI prediction using machine learning for children younger than 2 years (the most vulnerable age). This study aims to develop and evaluate ML models for the risk prediction of ARIs in children aged 6 months to 2 years. By addressing these objectives, the study aspires to contribute to the growing body of evidence on ML applications in pediatric healthcare, ultimately improving early diagnosis and treatment strategies for acute respiratory infections. 2. Methodology A secondary data analysis using 2016 Ethiopian Demographic and Health Survey (EDHS) data was done. In this study, the data were preprocessed and prepared, feature assessment was done, machine learning models were developed, the developed models were tested and deployed. Figure 1 shows the overall procedure of the study. 2.1. Data preparation and class balancing The demographic survey data collected from 2500 subjects, each having 24 variables, was obtained from the national repository. The 24 variables were maternal age, region, residence, mother’s education, religion, family size, sex of the household head, wealth index, currently pregnant, number of children, desire for more children, father’s education, father occupation, mother occupation, sex child, child age, ANC visit, place delivery, mode delivery, PNC visit, The variables included having diarrhea, experiencing a cough, being exposed to media, and having a specific micronutrient intake status. From these variables, this study aims to detect the presence of cough (had cough) variables from the information of the remaining 23 variables. The variable values were presented in the form of nominal, ordinal, and range. For convenience, while performing feature importance and machine learning tasks, the nominal, ordinal, and range values were represented using numbers. Table 1 shows the nominal, ordinal, and range values for the variables and their corresponding numerical representations as per their appearance order. The text describes a machine learning model that uses a combination of ordinal and binary categories, as well as categorical variables. Ordinal categories are represented by one hot encoding, while categorical variables are represented by numbers starting from the lowest number to the highest. Some variables are not ordinal but are binary categories with answers of yes or no, which are also considered ordinal. These variables are converted into categories by assigning 0 for no and 1 for yes. The approach increases the number of variables, so some variables can be updated to ordinal, some to variables with yes or no values, and others to ones using one hot encoding. For example, the variable residence can be updated to rural-residence, with answers for urban residents as no and yes for rural residents. Other variables can be updated to have yes or no values, such as family size, sex_househead, number of children, father occupation, mother occupation, sex_child, place delivery, and mode delivery. Region and religion are converted to their nominal values, as there is no relationship between the values. In summary, the machine learning model uses a combination of ordinal and categorical variables to implement a machine learning model. 2.2. Feature assessment After representing variable values using numbers, the second task was assessing the given data by calculating the feature importance. The feature importance entails how each feature is influencing the output. The feature importance shows how the change in a particular feature will cause the output value to change. The study calculates the weight of the variable's impact in causing cough. The logistic regression and random forest algorithms were used to figure out how important each feature was. The importance of each feature was then found by taking the average of the results from the two algorithms. The Shapley Additive Explanation (SHAP) technique is used to show how the features affect the desired output as well as figure out how important each feature is. For analyzing the features, the importance level and impact on the output feature importance calculation and SHAP plot were done. 2.3. Model development and testing The training was done with the one-dimensional convolutional neural network (1DCNN), random forest (RF), support vector machine (SVM), logistic regression (LR), gradient boost (GBC), K nearest neighbor (KNN), and decision tree (DT) algorithms after the data was cleaned up and made ready for use by machine learning (ML) algorithms. Data splitting was essential for performing the training activity and achieving a more stable result. Since the data was divided into training and testing groups, maybe an imbalance exists. At least one group will dominate the test or training sets, while the other will dominate the training set. Because of this, the outcomes of our machine learning model development will always be subject to bias and will vary from one instance of the train-test split to the next. The approach of stratified K-fold cross-validation was used to investigate and overcome these concerns. Every data point will be used in this manner for testing and training. Because of this, the data will be divided into folds or subsets that include the same percentage of each class level. This leads to production of performance that is superior in terms of precision, reliability, and consistency. To determine how successful a machine learning classifier is, several performance assessment methods may be applied. For evaluating the performance of each classification model, the metrics of precision, recall, F1 score, accuracy, and area under the curve (AUC) were used. The area under the receiver operating characteristic curve (AUC_ROC) score will be 1 for the ideal model. Models that are not superior to random guessing will get a score of 0.5, while models that are inferior to random guessing will receive a score that is lower than 0.5. To do a lot of parameters tuning, grid search, random search, and the optuna parameter tuning algorithm were used. When the process was over, the best set of hyperparameters was chosen. 1.1. 2.4. Deployment The best-performing machine learning model was selected based on the results of accuracy, precision, recall, and AUC performance evaluation metrics. Two widely used, accessible, and free platforms, GitHub ( https://github.com/ ) and Streamlit ( https://share.streamlit.io/ ), were selected for deploying the developed model as a sample ML backend web-based application. A free code repository and coding platform, GitHub, was used as a repository for the developed model and the graphical user interface code and requirement text, which are necessary for web app deployment. A free web app hosting platform, Streamlit, was used as a front-end interface where the users can interact with the developed and selected model for predicting the presence of ARI through a visually intuitive web app. 3. Results 3.1. Data processing Table 1: The initial variables with their given variables and the modified variables with their respective encodings. Given variable Initial values Modified variables Modified values Respective encodings maternal_age '15-24','25-34','>=35' ‘maternal_age’ '15-24','25-34','>=35' 0,1,2 Region 'tigray','afar','amhara', 'oromia','snnpr', 'benishangul','somali', 'harari','gambela', 'addis Abeba','dire dawa' 'tigray','afar','amhara', 'oromia','snnpr', 'benishangul','somali', 'harari','gambela', 'addis Abeba','dire dawa' No, yes 0, 1 Residence 'rural’, ‘urban’ ‘rural-residence’ No, yes 0,1 Mother_education 'no education', 'primary', 'secondary','higher' Mother_uneducated No, yes 0, 1 Religion 'Orthodox','Muslim', 'Protestant','Others' 'Orthodox','Muslim', 'Protestant','Others' No, yes 0,1 family_size '5' ‘family-size >5’ No, yes 0,1 sex_householdhead 'male','female' ‘male-househead’ No, yes 0,1 wealth_index 'Poor','Middle','Rich' ‘wealth_index’ 'Poor','Middle','Rich' 0,1,2 currently_pregnant 'no or unsure','yes' currently_pregnant 'no or unsure','yes' 0,1 number_children '=4' number of children > 3’ No, yes 0,1 desire_morechildren 'Wants no more', 'Undecided','Wants' desire_morechildren 'Wants no more', 'Undecided','Wants' 0,1,2 father_education 'no education', 'primary', 'secondary','higher' Father_uneducated No, yes 0,1 father_occupation 'Not working','Working' ‘father_unemployed’ No, yes 0, 1 Mother_occupation 'Not working','Working' Mother_ unemployed No, yes 0, 1 Sex_child 'male','female' ‘male-child-sex’ No, yes 0, 1 chid_age '6-11','12-23' chid_age '6-11','12-23' 0, 1 ANC_visit 'No','Yes' ANC_visit 'No','Yes' 0, 1 Place_delivery 'Home','Health facility' ‘home delivery’ 'No','Yes' 0, 1 mode_delivery 'Spontanous vaginal delivery','Cesarean section' 'cesarean section delivery' 'No','Yes' 0, 1 PNC_visit 'no','yes' PNC_visit 'no','yes' 0, 1 Had_diarrhea 'No','Yes' Had_diarrhea 'No','Yes' 0, 1 had_cough 'No','Yes' had_cough 'No','Yes' 0, 1 media_exposure 'No','Yes' media_exposure 'No','Yes' 0, 1 Micronutrient_intakestatus 'No','Yes' Micronutrient_intakestatus 'No','Yes' 0, 1 Among 2500 subjects, 503 had a cough, and the remaining 1997 subjects didn’t have one. This shows the data is highly imbalanced, and the results obtained and the models developed from it will be biased. So, synthetic minority oversampling techniques (SMOTE) were used to make more minority classes by interpolating between neighbors. This increased the number of instances of the minority class, which will help get rid of any possible bias in the machine learning model. The application of SMOTE led to an increase in the data to 3992, with 1996 elements in each class, including those with and without coughs. The tasks of evaluating features and building a machine learning model were done before and after the data balancing algorithm was used to see how the imbalance affected them. 3.2. Feature assessment result The feature importances were computed using logistic regression and random forest algorithms. Table 2 shows the results obtained from feature importance calculations. After applying SMOTE, the variables mode of delivery, region, and current pregnancy status resulted in 0.09, 0.089, and 0.079 values, respectively. Table 2: Feature importance results before and after applying SMOTE Feature Before SMOTE After SMOTE LR RF Mean LR RF Mean maternal_age 0.008 0.069 0.039 0.008 0.062 0.035 rural-residence 0.004 0.021 0.013 0.007 0.018 0.012 Mother_uneducated 0.008 0.036 0.022 0.005 0.031 0.018 family-size >5 0.008 0.039 0.024 0.001 0.033 0.017 male-househead 0.005 0.032 0.019 0.007 0.03 0.019 wealth_index 0.005 0.056 0.03 0.003 0.047 0.025 currently_pregnant 0.045 0.013 0.029 0.026 0.018 0.022 number of children > 3 0.016 0.031 0.023 0.004 0.025 0.015 desire_morechildren 0.002 0.051 0.026 0 0.044 0.022 Father_uneducated 0.015 0.038 0.027 0.008 0.034 0.021 father_unemployed 0.028 0.02 0.024 0.007 0.017 0.012 Mother_ unemployed 0.041 0.041 0.041 0.02 0.052 0.036 male-child-sex 0.026 0.052 0.039 0.002 0.045 0.023 chid_age 0.011 0.05 0.03 0.006 0.04 0.023 ANC_visit 0.001 0.034 0.018 0.011 0.033 0.022 home delivery 0.011 0.035 0.023 0.003 0.032 0.018 cesarean section delivery 0.056 0.008 0.032 0.031 0.012 0.022 PNC_visit 0.003 0.025 0.014 0.016 0.022 0.019 Had_diarrhea 0.116 0.058 0.087 0.018 0.034 0.026 media_exposure 0.029 0.038 0.033 0.003 0.033 0.018 Micronutrient_intakestatus 0.006 0.039 0.023 0.006 0.04 0.023 region_tigray 0.122 0.031 0.077 0.029 0.018 0.024 region_afar 0.035 0.013 0.024 0.046 0.018 0.032 region_amhara 0.025 0.014 0.02 0.053 0.021 0.037 region_oromia 0.042 0.021 0.032 0.047 0.023 0.035 region_somali 0.059 0.01 0.034 0.073 0.031 0.052 region_benishangul 0.108 0.013 0.061 0.096 0.044 0.07 region_snnpr 0.023 0.015 0.019 0.049 0.019 0.034 region_gambela 0.033 0.011 0.022 0.071 0.017 0.044 region_harari 0.066 0.01 0.038 0.082 0.027 0.054 region_addis adaba 0.008 0.007 0.007 0.045 0.01 0.027 region_dire dawa 0.011 0.012 0.011 0.056 0.014 0.035 religion_Muslim 0.006 0.017 0.012 0.034 0.018 0.026 religion_Orthodox 0.011 0.015 0.013 0.038 0.017 0.028 religion_Protestant 0.006 0.016 0.011 0.038 0.017 0.027 religion_Others 0.001 0.006 0.003 0.049 0.005 0.027 Figure 2 illustrates the SHAP bar plot before and after applying the SMOTE data balancing technique. Figure 3 demonstrates the most influential variables, and the variable region significantly influences the output. 3.2. Model development and testing results 1DCNN, RF, SVM, LR, GBC, KNN, and DT models were developed. The models were tested and got an AUC score of 0.81, 0.91, 0.77, 0.73, 0.72, 0.85, and 0.76, in that order, after the SMOTE data class balancing technique was used. Table 3 presents the performance results of the developed machine learning models before and after applying SMOTE. The performance results obtained before the application of the data class balancing technique were significantly lower than the results obtained after the application of data class balancing. This entails that data class balancing plays an essential role in overcoming high-level data class imbalance challenges while developing machine learning models. Based on the results obtained, the random forest model is the best performer with an AUC score of 0.91 and an accuracy of 0.83. Table 3. The models’ performance scores before and after applying data class balancing Model name Accuracy Recall F1_score Precision AUC-score Evaluation results before applying SMOTE LR 0.799 0.114 0.186 0.508 0.700 SVM 0.797 0.034 0.062 0.504 0.641 KNN 0.784 0.090 0.143 0.357 0.592 DT 0.701 0.292 0.282 0.275 0.548 RF 0.785 0.119 0.182 0.399 0.655 GBC 0.798 0 0 0 0.692 1DCNN 0.742 0.256 0.283 0.326 0.599 Evaluation results after applying SMOTE LR 0.758 0.745 0.754 0.764 0.842 SVM 0.804 0.790 0.801 0.813 0.881 KNN 0.749 0.914 0.784 0.687 0.860 DT 0.788 0.827 0.796 0.767 0.792 RF 0.840 0.862 0.844 0.827 0.918 GBC 0.668 0.716 0.683 0.658 0.726 1DCNN 0.800 0.824 0.804 0.788 0.872 As shown in Figure 4a, the ROC curve for the models using the data prior to SMOTE lies just above the straight line, meaning their performance is slightly better than random guessing but not very well. On the other hand, the ROC curves of the models developed using the balanced data given in Figure 4b showed improvements, and some of the models are performing very well. It is visible that the ROC curve of the random forest model is expanding to the left upper corner after applying data balancing. Figure 5 is a confusion matrix plot that shows how many wrongly classified data points there were in models that were made with unbalanced data (Figure 5a) and balanced data (Figure 5b). 3.3. Deployment The machine learning backend application, developed in the form of web-based app (accessible by the link https://aridetector.streamlit.app/) and desktop app (can be accessed from https://drive.google.com/file/d/1iwuyaVyTbqYhh3bfOAGifLjc77Vp98TD/view) for predicting the probability of cough presence. The online app can be used directly by using the link, while the desktop version requires to download and install on the local computer. In both online and desktop applications, the key features for determining the possibility of cough presence based on demographic information were prepared for the users to select their corresponding weights or nominal values. Users can select the weightage or set nominal values for each factor in the applications. The applications also include a decision button that allows users to predict the likelihood of a cough by clicking the submit button. Finally, the system will display the prediction result. Figure 6(a) shows the online app outline, while Figure 6(b) shows the desktop app frontend. 4. Discussion ARI is the leading cause of death and disability of children under the age of five years [ 1 , 2 ]. Malnutrition, inadequate immunization, poor sanitation, and limited access to healthcare facilities are among the main factors contributing to ARI in LMICs [ 1 , 2 , 3 ]. ARIs, especially pneumonia, are the major healthcare challenges in Ethiopia [ 12 ]. As traditional diagnostic and monitoring approaches are time-consuming and inaccessible to all early predictions, ML is becoming a more robust tool for early prediction [ 5 , 6 , 7 ]. Children 6 months to 2 years old are highly vulnerable to ARI, yet early prediction of ARI utilizing the robust ML algorithm is not explored [ 8 , 9 , 10 , 11 ]. The aim of this study is to develop ML models for predicting the ARI in children aged 6 months to 2 years. The data obtained from the national repository contains 2500 entities; out of those, 503 had a cough, and the remaining in 1997 didn’t have a cough. The ratio of subjects having a cough to those who didn’t is almost 1:4, making the data highly imbalanced. Using this imbalanced data, LR, SVM, KNN, DT, RF, GBC, and 1DCNN models were developed and resulted in AUC scores of 0.700, 0.641, 0.592, 0.548, 0.655, 0.692, and 0.599 and recall scores of 0.114, 0.034, 0.090, 0.292, 0.119, 0.00, and 0.256, respectively. The results presented lack clarity. As Data class balance is crucial for developing machine learning models; therefore, SMOTE data balancing techniques were employed, resulting in 3992 data elements with 1996 entities in both the 'had-cough' and 'didn’t-have-cough' classes. Using the balanced data, AUC scores of 0.842, 0.881, 0.860, 0.792, 0.918, 0.726, and 0.872 and recall scores of 0.745, 0.790, 0.914, 0.827, 0.862, 0.716, and 0.824 were obtained for LR, SVM, KNN, DT, RF, GBC, and 1DCNN models, respectively. Using k-fold cross-validation, initially, before the data-balancing operation, the average number of data allocated for testing was 500. Out of 500 data elements, 100, 101, 118, 149, 107, 100, and 129 were wrongly classified by LR, SVM, KNN, DT, RF, GBC, and 1DCNN models, respectively. The percentages of wrongly classified cases were 20%, 20.2%, 23.6%, 29.8%, 21.4%, 20%, and 25.8%. Upon balancing the data classes, the testing set increased to 798 data elements, and 193, 156, 200, 169, 127, 264, and 159 were wrongly classified by LR, SVM, KNN, DT, RF, GBC, and 1DCNN models, respectively. The percentages of wrongly classified cases were 24%, 19.5%, 25%, 21.2%, 15.9%, 33%, and 19.9%. These results show a significant improvement in performance for random forests and support vector machines in classifying each entity correctly. The findings show that SMOTE data balancing techniques have great importance for overcoming the bias and alteration in the development of ML models. Additionally, performance results entail that from the demographic survey data, it is possible to predict the possibility of ARIs using ML models from balanced data. A study done to develop a machine learning model from DHS data showed AUC scores of 0.79, 0.94, 0.73, 0.73, 0.90, 0.80, and 0.81 and sensitivities of 0.88, 0.87, 0.68, 0.72, 0.79, 0.68, and 0.77 for decision tree, random forest, Naïve Bayes, logistic regression, KNN, SVM, and gradient boosting models, respectively [ 18 ]. With a recall of 0.87 and an AUC score of 0.94, the Forest algorithm performed best in predicting ARI from DHS data [ 18 ]. A study showed a random forest ML model can predict ARI from DHS data with an accuracy of 0.96 [ 22 ]. An ensemble model was able to predict the ARI with an accuracy of 0.86 and sensitivity of 0.84 [ 23 ]. An AUC score of 0.95, a precision of 0.89, an accuracy of 0.88, an F-1 score of 0.83, and a recall of 0.77 were obtained from a random forest model trained with demographic health survey data [ 29 ]. Using the demographic health survey obtained in 2016 in the Ethiopian population, the XGBoost machine learning model was scoring an accuracy of 0.79 and an AUC score of 0.86 for predicting ARI [ 25 ]. From these results, the RF model demonstrates a significant improvement in classification accuracy due to data-class balancing. Thus, the RF model was selected for GUI development. The developed GUI enables the users to select the nominal, ordinal, or range values for the given demographic health survey (DHS) data collection and predicts the percentage of a certain DHS entry to be classified as having a cough or not having a cough. Developing an online and desktop application from the best-performing machine learning model enables it to be evaluated thoroughly by the researchers, health care service providers, and regulatory authorities. These evaluations will make a great contribution to further improvement of the model and the GUI system. Then, finally, upon evaluation of the developed system by the regulatory authorities, this system can be used to predict the potential forthcoming cough from demographic health survey data. Researchers were able to create and utilize machine learning applications for healthcare that work online and offline. These applications can predict if someone has COVID-19 based on user input [ 30 ], manage electronic health records for long-term conditions [ 31 ], visualize bioinformatics and cheminformatics projects [ 32 ], predict cancer, heart diseases, diabetes, and kidney diseases using the Streamlit web hosting platform [ 33 ], and estimate child deaths from pneumonia[ 34 ]. Conclusion This study utilized nationally representative Demographic and Health Survey (DHS) data to develop machine learning models for predicting acute respiratory infections (ARI) in children. Algorithms implemented included logistic regression, support vector machines, k-nearest neighbors, decision trees, random forests, gradient boosting classifiers, and one-dimensional convolutional neural networks (1D-CNN). Feature importance was assessed using random forest, logistic regression, and SHAP (Shapley Additive Explanations), identifying region, maternal unemployment, and paternal educational status as key predictors of ARI. Among all models, the random forest algorithm demonstrated superior performance based on accuracy, recall, and area under the curve (AUC) metrics. This best-performing model was subsequently deployed as a web application via Streamlit, with an offline version also prepared for Windows systems. These findings demonstrate the feasibility of leveraging demographic data and machine learning for early ARI prediction, particularly in resource-constrained settings. Declarations Ethics approval and consent to participate This study is a secondary data analysis from the EDHS data, so it does not require ethical approval. For conducting this study, online registration and requests for measuring DHS were conducted. The dataset was downloaded from the DHS online archive from the MEASURE DHS dataset for free after getting approval to access the data. Patient and public involvement The study was conducted using secondary data and did not involve patient or public participation. Consent for publication Not applicable. Availability of data and material The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request. Competing interests The authors have no conflicts of interests Funding The authors did not receive any funds for publication Authors’ contribution All authors have been involved in all steps from conceptualization, data extraction, analysis, and manuscript writing. All authors proofread the manuscript and approved its submission for publication. Acknowledgments Not applicable References M. Cashat-Cruz, J. J. Morales-Aguirre, and M. Mendoza-Azpiri, “Respiratory tract infections in children in developing countries,” Semin. Pediatr. Infect. Dis. , vol. 16, no. 2, pp. 84–92, Apr. 2005, doi: 10.1053/j.spid.2005.12.005. “Pneumonia in children.” Accessed: Mar. 12, 2025. [Online]. Available: https://www.who.int/news-room/fact-sheets/detail/pneumonia J. O. Akinyemi and O. M. Morakinyo, “Household environment and symptoms of childhood acute respiratory tract infections in Nigeria, 2003–2013: a decade of progress and stagnation,” BMC Infect. Dis. , vol. 18, no. 1, p. 296, July 2018, doi: 10.1186/s12879-018-3207-5. S. Hassen et al. , “Determinants of acute respiratory infection (ARI) among under-five children in rural areas of Legambo District, South Wollo Zone, Ethiopia: A matched case–control study,” Int. J. Infect. Dis. , vol. 96, pp. 688–695, July 2020, doi: 10.1016/j.ijid.2020.05.012. Faculty of Public Health, Universitas Indonesia and E. Triana, “Factors Affecting The Incidence of Acute Respiratory Tract Infection in Children under Five at Betungan Community Health Center, Bengkulu,” in Strengthening Hospital Competitiveness to Improve Patient Satisfaction and Better Health Outcomes , Masters Program in Public Health, Graduate School, Universitas Sebelas Maret, 2019, pp. 40–45. doi: 10.26911/the6thicph-FP.01.06. M. Sultana et al. , “Prevalence, determinants and health care-seeking behavior of childhood acute respiratory tract infections in Bangladesh,” PLOS ONE , vol. 14, no. 1, p. e0210433, Jan. 2019, doi: 10.1371/journal.pone.0210433. I. Mejía-Guevara, W. Zuo, E. Bendavid, N. Li, and S. Tuljapurkar, “Age distribution, trends, and forecasts of under-5 mortality in 31 sub-Saharan African countries: A modeling study,” PLOS Med. , vol. 16, no. 3, p. e1002757, Mar. 2019, doi: 10.1371/journal.pmed.1002757. S. Yaya, G. Bishwajit, F. Okonofua, and O. A. Uthman, “Under five mortality patterns and associated maternal risk factors in sub-Saharan Africa: A multi-country analysis,” PLOS ONE , vol. 13, no. 10, p. e0205977, Oct. 2018, doi: 10.1371/journal.pone.0205977. M. Sonego, M. C. Pellegrin, G. Becker, and M. Lazzerini, “Risk Factors for Mortality from Acute Lower Respiratory Infections (ALRI) in Children under Five Years of Age in Low and Middle-Income Countries: A Systematic Review and Meta-Analysis of Observational Studies,” PLOS ONE , vol. 10, no. 1, p. e0116380, Jan. 2015, doi: 10.1371/journal.pone.0116380. E. Díaz-Martínez and E. D. and Gibbons, “The Questionable Power of the Millennium Development Goal to Reduce Child Mortality,” J. Hum. Dev. Capab. , vol. 15, no. 2–3, pp. 203–217, July 2014, doi: 10.1080/19452829.2013.864621. A. Prüss-Ustün et al. , “Burden of disease from inadequate water, sanitation and hygiene for selected adverse health outcomes: An updated analysis with a focus on low- and middle-income countries,” Int. J. Hyg. Environ. Health , vol. 222, no. 5, pp. 765–777, June 2019, doi: 10.1016/j.ijheh.2019.05.004. A. Keleb et al. , “Pneumonia remains a leading public health problem among under-five children in peri-urban areas of north-eastern Ethiopia,” PLOS ONE , vol. 15, no. 9, p. e0235818, Sept. 2020, doi: 10.1371/journal.pone.0235818. T. Nigatu et al. , “The status of immunization program and challenges in Ethiopia: A mixed method study,” SAGE Open Med. , vol. 12, p. 20503121241237115, June 2024, doi: 10.1177/20503121241237115. F. Epelde, “How AI Could Help Us in the Epidemiology and Diagnosis of Acute Respiratory Infections?,” Pathogens , vol. 13, no. 11, Art. no. 11, Nov. 2024, doi: 10.3390/pathogens13110940. D. A. Rankin et al. , “Diagnostic models predicting paediatric viral acute respiratory infections: a systematic review,” BMJ Open , vol. 13, no. 4, p. e067878, Apr. 2023, doi: 10.1136/bmjopen-2022-067878. P. Yadav, V. Rastogi, A. Yadav, and P. Parashar, “Artificial Intelligence: A promising tool in diagnosis of respiratory diseases,” Intell. Pharm. , vol. 2, no. 6, pp. 784–791, Dec. 2024, doi: 10.1016/j.ipha.2024.05.002. S. Al-Anazi et al. , “Artificial intelligence in respiratory care: Current scenario and future perspective,” Ann. Thorac. Med. , vol. 19, no. 2, p. 117, June 2024, doi: 10.4103/atm.atm_192_23. A. K. Kassaw, G. Bekele, A. K. Kassaw, and A. Yimer, “Prediction of acute respiratory infections using machine learning techniques in Amhara Region, Ethiopia,” Sci. Rep. , vol. 14, no. 1, p. 27968, Nov. 2024, doi: 10.1038/s41598-024-76847-3. H. M. Fenta, T. T. Zewotir, S. Naidoo, R. N. Naidoo, and H. Mwambi, “Factors of acute respiratory infection among under-five children across sub-Saharan African countries using machine learning approaches,” Sci. Rep. , vol. 14, no. 1, p. 15801, July 2024, doi: 10.1038/s41598-024-65620-1. Y. Ku, S. B. Kwon, J.-H. Yoon, S.-K. Mun, and M. Chang, “Machine Learning Models for Predicting the Occurrence of Respiratory Diseases Using Climatic and Air-Pollution Factors,” Clin. Exp. Otorhinolaryngol. , vol. 15, no. 2, pp. 168–176, Jan. 2022, doi: 10.21053/ceo.2021.01536. R. M. Kananura, “Machine learning predictive modelling for identification of predictors of acute respiratory infection and diarrhoea in Uganda’s rural and urban settings,” PLOS Glob. Public Health , vol. 2, no. 5, p. e0000430, May 2022, doi: 10.1371/journal.pgph.0000430. T. Z. Yehuala, B. M. Fente, S. M. Wubante, and N. M. Derseh, “Exploring machine learning algorithms to predict acute respiratory tract infection and identify its determinants among children under five in Sub-Saharan Africa,” Front. Pediatr. , vol. 12, Nov. 2024, doi: 10.3389/fped.2024.1388820. M. H. Kalayou, A.-A. K. Kassaw, and K. B. Shiferaw, “Empowering child health: Harnessing machine learning to predict acute respiratory infections in Ethiopian under-fives using demographic and health survey insights,” BMC Infect. Dis. , vol. 24, no. 1, p. 338, Mar. 2024, doi: 10.1186/s12879-024-09195-2. T.-H. Chang et al. , “Clinical characteristics of hospitalized children with community-acquired pneumonia and respiratory infections: Using machine learning approaches to support pathogen prediction at admission,” J. Microbiol. Immunol. Infect. , vol. 56, no. 4, pp. 772–781, Aug. 2023, doi: 10.1016/j.jmii.2023.04.011. Z. B. Tadese et al. , “Interpretable prediction of acute respiratory infection disease among under-five children in Ethiopia using ensemble machine learning and Shapley additive explanations (SHAP),” Digit. Health , vol. 10, p. 20552076241272739, Sept. 2024, doi: 10.1177/20552076241272739. “Leveraging Machine Learning for Predictive Models in Healthcare to Enhance Patient Outcome Management,” Int. Res. J. Mod. Eng. Technol. Sci. , Jan. 2025, doi: 10.56726/IRJMETS66198. A. Rahman, M. Karmakar, and P. Debnath, “Predictive Analytics for Healthcare: Improving Patient Outcomes in the U.S. through Machine Learning,” Rev. Intel. Artif. En Med. , vol. 14, no. 1, Art. no. 1, Nov. 2023. H. Ali, “AI for Pandemic Preparedness and Infectious Disease Surveillance: Predicting Outbreaks, Modeling Transmission, and Optimizing Public Health Interventions,” Int. J. Res. Publ. Rev. , vol. 6, no. 2, pp. 4605–4619, Feb. 2025, doi: 10.55248/gengpi.6.0225.0941. T. Z. Yehuala et al. , “Machine learning algorithms to predict healthcare-seeking behaviors of mothers for acute respiratory infections and their determinants among children under five in sub-Saharan Africa,” Front. Public Health , vol. 12, June 2024, doi: 10.3389/fpubh.2024.1362392. C. N. Villavicencio, J. J. Macrohon, X. A. Inbaraj, J.-H. Jeng, and J.-G. Hsieh, “Development of a Machine Learning Based Web Application for Early Diagnosis of COVID-19 Based on Symptoms,” Diagnostics , vol. 12, no. 4, Art. no. 4, Apr. 2022, doi: 10.3390/diagnostics12040821. S. Burns, A. Cushing, A. Taylor, D. J. Lowe, and C. Carlin, “Supporting long-term condition management: a workflow framework for the co-development and operationalization of machine learning models using electronic health record data insights,” Front. Artif. Intell. , vol. 7, p. 1458508, Nov. 2024, doi: 10.3389/frai.2024.1458508. C. Nantasenamat, A. Biswas, J. M. Napoles, M. Parker, and R. Dunbrack, “Building bioinformatics web applications with Streamlit,” 2023, pp. 679–699. doi: 10.1016/B978-0-443-18638-7.00001-3. L. D. Gopisetti, S. K. L. Kummera, S. R. Pattamsetti, S. Kuna, N. Parsi, and H. P. Kodali, “Multiple Disease Prediction System using Machine Learning and Streamlit,” in 2023 5th International Conference on Smart Systems and Inventive Technology (ICSSIT) , Jan. 2023, pp. 923–931. doi: 10.1109/ICSSIT55814.2023.10060903. N. I. Mohammed, A. Jarde, G. Mackenzie, U. D’Alessandro, and D. Jeffries, “Deploying Machine Learning Models Using Progressive Web Applications: Implementation Using a Neural Network Prediction Model for Pneumonia Related Child Mortality in The Gambia,” Front. Public Health , vol. 9, p. 772620, Feb. 2022, doi: 10.3389/fpubh.2021.772620. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 04 Oct, 2025 Reviewers agreed at journal 27 Sep, 2025 Reviews received at journal 26 Sep, 2025 Reviewers agreed at journal 25 Sep, 2025 Reviewers invited by journal 18 Sep, 2025 Editor invited by journal 25 Aug, 2025 Editor assigned by journal 22 Aug, 2025 Submission checks completed at journal 22 Aug, 2025 First submitted to journal 21 Aug, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7425776","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":546766039,"identity":"876f8d8c-194d-42d9-abf4-7a6257aeb23f","order_by":0,"name":"Ewunate Assaye Kassaw","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4ElEQVRIiWNgGAWjYPCC/3b87A1A2sCCaC3MyZI9B0BaJIjXwrjhRgKIQYQW/v7jDz/zVLAxS858fnXDjwIJBv727gS8WiRu5BhL85zh4eOXzim72QN0mMSZsxvwW3ODh0Gat02CWXJ2TtoNHqAWA4lc/Frkzx9//Jv3nwHjhptn0m7+IUaLwYEEM2nehgSg99mP3SbKFsMbOWaWc44dAAZyDtttGQMJHoJ+kQM67MabmgPAqDz+7OabPzZy/O29BLwPBEw8YIrHAEwSVA4CjD/AFPsDolSPglEwCkbByAMAw0FIHgmcr3sAAAAASUVORK5CYII=","orcid":"","institution":"University of Gondar","correspondingAuthor":true,"prefix":"","firstName":"Ewunate","middleName":"Assaye","lastName":"Kassaw","suffix":""},{"id":546766041,"identity":"ab48c01f-c74d-426a-8fd2-6b4faeaaf200","order_by":1,"name":"Biruk Beletew Abate","email":"","orcid":"","institution":"Woldia University","correspondingAuthor":false,"prefix":"","firstName":"Biruk","middleName":"Beletew","lastName":"Abate","suffix":""},{"id":546766042,"identity":"1026545f-03d3-4c52-8d16-ae061b0d029f","order_by":2,"name":"Ashenafi Kibret Sendekie","email":"","orcid":"","institution":"University of Gondar","correspondingAuthor":false,"prefix":"","firstName":"Ashenafi","middleName":"Kibret","lastName":"Sendekie","suffix":""}],"badges":[],"createdAt":"2025-08-21 11:38:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7425776/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7425776/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97794120,"identity":"8877b6a4-c133-4d65-a1bb-922b1de19192","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":654501,"visible":true,"origin":"","legend":"","description":"","filename":"AcuterespiratoryinfectionsriskpredictionusingmachinelearningamongEthiopianchildrenAged6Monthsto2Years.docx","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/aa5252318499a6a3077fe5df.docx"},{"id":97898782,"identity":"9569c424-782b-4a7e-b5ab-a3ce8d7623f9","added_by":"auto","created_at":"2025-12-10 15:39:38","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6462,"visible":true,"origin":"","legend":"","description":"","filename":"45e65a959c77476ca90fc8f6de97482b.json","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/0838cd77473b53e42c9ffd4d.json"},{"id":97794122,"identity":"0ca534b5-ece1-4807-a42f-f111d160a71e","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":104045,"visible":true,"origin":"","legend":"","description":"","filename":"45e65a959c77476ca90fc8f6de97482b1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/f4ef3b045bc89c2cd0c5097e.xml"},{"id":97794142,"identity":"6343ac29-42f7-479b-8cb9-ef5c2549a91b","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"jpeg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":142400,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/15e48163fcacb138137e17ce.jpeg"},{"id":97898662,"identity":"5c4a07fd-40b3-482a-9065-76887c72b5dc","added_by":"auto","created_at":"2025-12-10 15:39:26","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":116560,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/951de72271f3606545d76045.png"},{"id":97895934,"identity":"dc7f8ab9-a46d-484d-8c73-91e7adf56378","added_by":"auto","created_at":"2025-12-10 15:35:25","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22305,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/148cad350e48989577ac4765.png"},{"id":97897362,"identity":"407e8488-1c8d-4af8-b4c0-e8957815702a","added_by":"auto","created_at":"2025-12-10 15:37:46","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":24022,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/85e23143c269a8032705d59b.png"},{"id":97794134,"identity":"c0d993c5-68be-471d-b39a-4e47b09dafc9","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":45487,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/f1007631367acf5730309475.png"},{"id":97794128,"identity":"193d55d7-557e-47e6-9c5c-d14319661f22","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":41613,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/2da63bdfede4eb8461208200.png"},{"id":97898305,"identity":"5db6e0bb-9cf6-42ab-bbac-a1ae0e5e0143","added_by":"auto","created_at":"2025-12-10 15:38:59","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":66945,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/4fe8bbb14eb6d7dd8fb5d4b5.jpeg"},{"id":97794132,"identity":"6a692767-1d5f-413f-ad46-c9318cdb67bd","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":74754,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/6aead909e0feea06747a4524.png"},{"id":97898693,"identity":"315013dd-2888-4179-a6d6-66273b88897b","added_by":"auto","created_at":"2025-12-10 15:39:28","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":72778,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/ba1eef5f99003a8f97834270.png"},{"id":97896075,"identity":"0bdc4512-50fd-46e3-96e1-d0dbcbf2600b","added_by":"auto","created_at":"2025-12-10 15:35:49","extension":"jpeg","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":381313,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/294ea90cd7c8d427fb9a2742.jpeg"},{"id":97897399,"identity":"0dd770d4-6052-4691-9dad-8f5b2883b689","added_by":"auto","created_at":"2025-12-10 15:37:48","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30356,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/f98b2cf7ad19c5ea8a8be937.png"},{"id":97897510,"identity":"e01d1e0a-e4d7-4f63-9dac-5235b2a97604","added_by":"auto","created_at":"2025-12-10 15:37:55","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":21089,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/6952a1baa42b8bdeba443c7b.png"},{"id":97794124,"identity":"302930c1-d93a-48f9-a7c5-18214705cbfc","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9312,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/b354abe0e9cec8d060afeff6.png"},{"id":97897220,"identity":"b2f4f73b-2cef-4a18-af47-76922fd84219","added_by":"auto","created_at":"2025-12-10 15:37:34","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9762,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/d26cc9ce1995a8b1d90d207e.png"},{"id":97895932,"identity":"7ead3fd3-a5b8-462d-88f8-85a8bb39a1fe","added_by":"auto","created_at":"2025-12-10 15:35:24","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16094,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/0de574ea55ad60f719434b0e.png"},{"id":97897417,"identity":"5a80a4b1-f409-4657-b729-8ae6d08aad83","added_by":"auto","created_at":"2025-12-10 15:37:48","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16642,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/deccdbe07530a491f0a11fd2.png"},{"id":97898339,"identity":"e80c7761-e5d2-4585-ab23-0058fbab985e","added_by":"auto","created_at":"2025-12-10 15:39:02","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13305,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/0a494154b9d704075f0a8bce.png"},{"id":97897495,"identity":"66727fed-5c52-4af0-a7a6-eef1cbc649f1","added_by":"auto","created_at":"2025-12-10 15:37:52","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22549,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/83a812e20965e38523925acf.png"},{"id":97794145,"identity":"1714831b-1002-4026-87a1-c1a68d0c4a09","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20975,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/0a5999bdcc421bb32295b9c5.png"},{"id":97896917,"identity":"6983b3e7-b2ae-49cc-90b9-252ff698465b","added_by":"auto","created_at":"2025-12-10 15:37:12","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":135405,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/5052978bb4bb9def6956f158.png"},{"id":97794148,"identity":"c790100e-e7a0-4dc8-b670-46b15e5bea23","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"xml","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102438,"visible":true,"origin":"","legend":"","description":"","filename":"45e65a959c77476ca90fc8f6de97482b1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/91a952c1a1d3eaa567160e08.xml"},{"id":97794147,"identity":"8c68427e-6da0-4267-a5b1-313fb4c71386","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"html","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":112195,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/42d120be52faed429229796c.html"},{"id":97794117,"identity":"f61376d2-db0f-4fc5-8255-b54e3e6bda01","added_by":"auto","created_at":"2025-12-09 12:34:02","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":15199,"visible":true,"origin":"","legend":"\u003cp\u003ethe overall procedure of the study.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/205db4c01635361c4b453788.png"},{"id":97794118,"identity":"7425c972-a406-4c7d-9f37-a87b7cf43e56","added_by":"auto","created_at":"2025-12-09 12:34:02","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":66259,"visible":true,"origin":"","legend":"\u003cp\u003ethe contribution weight of data features using the SHAP bar plot (a) before applying SMOTE and (b) after applying SMOTE.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/c4826f14640465734e16a3ed.png"},{"id":97794119,"identity":"8393b5b7-1d20-40f3-b927-3443c7d70755","added_by":"auto","created_at":"2025-12-09 12:34:02","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":113712,"visible":true,"origin":"","legend":"\u003cp\u003eThe SHAP summary plot before and after the data-class balancing\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/450621ed95ee6c95ec1d8b8e.png"},{"id":97794129,"identity":"32dade05-98be-499e-b8eb-fcf4d9ed84f8","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":103184,"visible":true,"origin":"","legend":"\u003cp\u003eThe ROC curves of the developed models (a) before applying SMOTE and (b) after applying SMOTE.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/1cb34ff4d7be266ecb2630c7.png"},{"id":97794139,"identity":"785011b5-22d3-4b39-9202-3ec7fd184c1d","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":111256,"visible":true,"origin":"","legend":"\u003cp\u003ethe confusion matrix of the developed models (a) before applying SMOTE and (b) after applying SMOTE.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/dd51f214d7ee01d1144f21cb.png"},{"id":97794123,"identity":"2e7fd76c-bae5-44b6-9514-571fcde7c2fb","added_by":"auto","created_at":"2025-12-09 12:34:03","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":248531,"visible":true,"origin":"","legend":"\u003cp\u003ethe developed graphical user interface system for predicting the probability of cough (a) the online app front end, (b) the desktop app front end of.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/f97aea7282217aab7b86d05a.png"},{"id":98797543,"identity":"836989db-d687-45bf-8107-1a267a25d8dc","added_by":"auto","created_at":"2025-12-22 13:13:17","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1374552,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7425776/v1/4a0ae8ea-42bd-4120-8404-09ba7074c617.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Acute respiratory infections risk prediction using machine learning among Ethiopian children Aged 6 Months to 2 Years","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eAcute respiratory infections (ARIs) are a group of diseases caused by a wide range of pathogens, including Streptococcus pneumoniae, Haemophilus influenzae, and Staphylococcus aureus.[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. ARIs are a leading cause of illness and death in children worldwide[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e], accounting for approximately 12\u0026nbsp;million cases and 1.3\u0026nbsp;million deaths annually among those under five years old[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. ARIs are particularly prevalent in low- and middle-income countries (LMICs), where they remain among the top causes of childhood morbidity and mortality[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e], [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e], [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. A significant proportion of these deaths occur in Sub-Saharan Africa and other developing regions[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Contributing factors in LMICs include limited access to healthcare, malnutrition, poor sanitation, and inadequate immunization coverage [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eEthiopia is among the top 15 countries with the highest burden of acute respiratory infections (ARIs), affecting approximately four million children each year[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. ARIs cause over 40,000 deaths annually among children under five, accounting for 18% of all child deaths in the country[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Pneumonia, a major form of ARI, contributes significantly to this mortality, driven by factors such as poverty, environmental pollution, and limited healthcare infrastructure[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Despite efforts to reduce mortality through vaccination programs and improved healthcare access, ARIs remain a major public health challenge in Ethiopia and similar settings[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eEarly and accurate prediction of ARIs is crucial for timely intervention and improved health outcomes[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Traditional diagnostic approaches rely on clinical assessment and laboratory tests, which may be time-consuming, costly, or inaccessible in resource-limited settings [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. In recent years, machine learning (ML) has emerged as a promising tool for predicting and diagnosing various health conditions, including respiratory diseases, by leveraging patient data and identifying patterns that may not be apparent through conventional analysis[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e], [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e].\u003c/p\u003e\u003cp\u003ePrevious studies have explored the application of ML in respiratory disease prediction, mainly focusing on older children and adults[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e], [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Recent studies have also predicted the risk factors of ARIs in children using ML approaches[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e], [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e], [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. The most employed ML algorithms include Random Forest (RF), Extreme Gradient Boosting (XGBoost), Gradient Boosting, Na\u0026iuml;ve Bayes (NB), Logistic Regression (LR), and Decision Trees, and among these, ensemble models\u0026mdash;particularly those combining RF [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] and XGBoost [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e] have consistently demonstrated superior performance.\u003c/p\u003e\u003cp\u003eAcross the literature, several recurring predictors of ARIs have been identified. At the child level, key factors include duration of breastfeeding, history of diarrhea, immunization status, and nutritional status[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Maternal and household variables such as educational attainment, occupation, wealth status, residential setting (urban or rural), cooking fuel, and the number of living children have also been associated with ARI risk[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eDespite these advancements, the existing literature lacks age-specific modelling tailored to infants and young children aged 6 to 24 months, who represent a particularly vulnerable subgroup due to their developmental stage and immature immune systems. Most existing studies have focused broadly on children under five, thus overlooking the unique risk profiles of this younger under two years. The rationale for this study stems from the need to develop an early prediction system that can aid healthcare providers in identifying at-risk children before severe complications arise. Given the challenges of early clinical diagnosis in infants and toddlers, who may present with non-specific symptoms, ML can be a valuable decision-support tool. ML models can offer high predictive accuracy by utilizing a combination of clinical, demographic, and environmental factors, assisting in timely clinical decision-making and reducing unnecessary hospital admissions[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Moreover, integrating such predictive models into community healthcare settings could enhance surveillance efforts and improve resource allocation in pediatric care[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eUnbalanced use of demographic health survey data, inability to consider the difference in the regional location of the subjects in the country, considering only infected subjects, inability to predict early, and inability to prepare the developed model into a utility tool were the research gaps of the existing studies done on the prediction of ARI using machine learning for the Ethiopian population. Additionally, researchers have not yet studied ARI prediction using machine learning for children younger than 2 years (the most vulnerable age).\u003c/p\u003e\u003cp\u003eThis study aims to develop and evaluate ML models for the risk prediction of ARIs in children aged 6 months to 2 years. By addressing these objectives, the study aspires to contribute to the growing body of evidence on ML applications in pediatric healthcare, ultimately improving early diagnosis and treatment strategies for acute respiratory infections.\u003c/p\u003e"},{"header":"2. Methodology","content":"\u003cp\u003eA secondary data analysis using 2016 Ethiopian Demographic and Health Survey (EDHS) data was done. In this study, the data were preprocessed and prepared, feature assessment was done, machine learning models were developed, the developed models were tested and deployed. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows the overall procedure of the study.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1. Data preparation and class balancing\u003c/h2\u003e\u003cp\u003eThe demographic survey data collected from 2500 subjects, each having 24 variables, was obtained from the national repository. The 24 variables were maternal age, region, residence, mother\u0026rsquo;s education, religion, family size, sex of the household head, wealth index, currently pregnant, number of children, desire for more children, father\u0026rsquo;s education, father occupation, mother occupation, sex child, child age, ANC visit, place delivery, mode delivery, PNC visit, The variables included having diarrhea, experiencing a cough, being exposed to media, and having a specific micronutrient intake status. From these variables, this study aims to detect the presence of cough (had cough) variables from the information of the remaining 23 variables. The variable values were presented in the form of nominal, ordinal, and range. For convenience, while performing feature importance and machine learning tasks, the nominal, ordinal, and range values were represented using numbers. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows the nominal, ordinal, and range values for the variables and their corresponding numerical representations as per their appearance order.\u003c/p\u003e\u003cp\u003eThe text describes a machine learning model that uses a combination of ordinal and binary categories, as well as categorical variables. Ordinal categories are represented by one hot encoding, while categorical variables are represented by numbers starting from the lowest number to the highest. Some variables are not ordinal but are binary categories with answers of yes or no, which are also considered ordinal. These variables are converted into categories by assigning 0 for no and 1 for yes. The approach increases the number of variables, so some variables can be updated to ordinal, some to variables with yes or no values, and others to ones using one hot encoding. For example, the variable residence can be updated to rural-residence, with answers for urban residents as no and yes for rural residents. Other variables can be updated to have yes or no values, such as family size, sex_househead, number of children, father occupation, mother occupation, sex_child, place delivery, and mode delivery. Region and religion are converted to their nominal values, as there is no relationship between the values. In summary, the machine learning model uses a combination of ordinal and categorical variables to implement a machine learning model.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2. Feature assessment\u003c/h2\u003e\u003cp\u003eAfter representing variable values using numbers, the second task was assessing the given data by calculating the feature importance. The feature importance entails how each feature is influencing the output. The feature importance shows how the change in a particular feature will cause the output value to change. The study calculates the weight of the variable's impact in causing cough. The logistic regression and random forest algorithms were used to figure out how important each feature was. The importance of each feature was then found by taking the average of the results from the two algorithms.\u003c/p\u003e\u003cp\u003eThe Shapley Additive Explanation (SHAP) technique is used to show how the features affect the desired output as well as figure out how important each feature is. For analyzing the features, the importance level and impact on the output feature importance calculation and SHAP plot were done.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3. Model development and testing\u003c/h2\u003e\u003cp\u003eThe training was done with the one-dimensional convolutional neural network (1DCNN), random forest (RF), support vector machine (SVM), logistic regression (LR), gradient boost (GBC), K nearest neighbor (KNN), and decision tree (DT) algorithms after the data was cleaned up and made ready for use by machine learning (ML) algorithms. Data splitting was essential for performing the training activity and achieving a more stable result.\u003c/p\u003e\u003cp\u003eSince the data was divided into training and testing groups, maybe an imbalance exists. At least one group will dominate the test or training sets, while the other will dominate the training set. Because of this, the outcomes of our machine learning model development will always be subject to bias and will vary from one instance of the train-test split to the next. The approach of stratified K-fold cross-validation was used to investigate and overcome these concerns. Every data point will be used in this manner for testing and training. Because of this, the data will be divided into folds or subsets that include the same percentage of each class level. This leads to production of performance that is superior in terms of precision, reliability, and consistency.\u003c/p\u003e\u003cp\u003eTo determine how successful a machine learning classifier is, several performance assessment methods may be applied. For evaluating the performance of each classification model, the metrics of precision, recall, F1 score, accuracy, and area under the curve (AUC) were used. The area under the receiver operating characteristic curve (AUC_ROC) score will be 1 for the ideal model. Models that are not superior to random guessing will get a score of 0.5, while models that are inferior to random guessing will receive a score that is lower than 0.5. To do a lot of parameters tuning, grid search, random search, and the optuna parameter tuning algorithm were used. When the process was over, the best set of hyperparameters was chosen.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e1.1. 2.4. Deployment\u003c/h2\u003e\u003cp\u003eThe best-performing machine learning model was selected based on the results of accuracy, precision, recall, and AUC performance evaluation metrics. Two widely used, accessible, and free platforms, GitHub (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/\u003c/span\u003e\u003cspan address=\"https://github.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and Streamlit (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://share.streamlit.io/\u003c/span\u003e\u003cspan address=\"https://share.streamlit.io/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), were selected for deploying the developed model as a sample ML backend web-based application. A free code repository and coding platform, GitHub, was used as a repository for the developed model and the graphical user interface code and requirement text, which are necessary for web app deployment. A free web app hosting platform, Streamlit, was used as a front-end interface where the users can interact with the developed and selected model for predicting the presence of ARI through a visually intuitive web app.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results","content":"\u003cp\u003e3.1.\u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Data processing\u003c/p\u003e\n\u003cp\u003eTable 1: The initial variables with their given variables and the modified variables with their respective encodings.\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"678\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003eGiven variable\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eInitial values\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003eModified variables\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eModified values\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003eRespective encodings\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003ematernal_age\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;15-24\u0026apos;,\u0026apos;25-34\u0026apos;,\u0026apos;\u0026gt;=35\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026lsquo;maternal_age\u0026rsquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;15-24\u0026apos;,\u0026apos;25-34\u0026apos;,\u0026apos;\u0026gt;=35\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0,1,2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003eRegion\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;tigray\u0026apos;,\u0026apos;afar\u0026apos;,\u0026apos;amhara\u0026apos;, \u0026apos;oromia\u0026apos;,\u0026apos;snnpr\u0026apos;, \u0026apos;benishangul\u0026apos;,\u0026apos;somali\u0026apos;, \u0026apos;harari\u0026apos;,\u0026apos;gambela\u0026apos;, \u0026apos;addis Abeba\u0026apos;,\u0026apos;dire dawa\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026apos;tigray\u0026apos;,\u0026apos;afar\u0026apos;,\u0026apos;amhara\u0026apos;, \u0026apos;oromia\u0026apos;,\u0026apos;snnpr\u0026apos;, \u0026apos;benishangul\u0026apos;,\u0026apos;somali\u0026apos;, \u0026apos;harari\u0026apos;,\u0026apos;gambela\u0026apos;, \u0026apos;addis Abeba\u0026apos;,\u0026apos;dire dawa\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003eResidence\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;rural\u0026rsquo;, \u0026lsquo;urban\u0026rsquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026lsquo;rural-residence\u0026rsquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0,1\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003eMother_education\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;no education\u0026apos;, \u0026apos;primary\u0026apos;, \u0026apos;secondary\u0026apos;,\u0026apos;higher\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003eMother_uneducated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003eReligion\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;Orthodox\u0026apos;,\u0026apos;Muslim\u0026apos;, \u0026apos;Protestant\u0026apos;,\u0026apos;Others\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026apos;Orthodox\u0026apos;,\u0026apos;Muslim\u0026apos;, \u0026apos;Protestant\u0026apos;,\u0026apos;Others\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0,1\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003efamily_size\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;\u0026lt;=5\u0026apos;,\u0026apos;\u0026gt;5\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026lsquo;family-size \u0026gt;5\u0026rsquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0,1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003esex_householdhead\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;male\u0026apos;,\u0026apos;female\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026lsquo;male-househead\u0026rsquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0,1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003ewealth_index\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;Poor\u0026apos;,\u0026apos;Middle\u0026apos;,\u0026apos;Rich\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026lsquo;wealth_index\u0026rsquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;Poor\u0026apos;,\u0026apos;Middle\u0026apos;,\u0026apos;Rich\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0,1,2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003ecurrently_pregnant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;no or unsure\u0026apos;,\u0026apos;yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003ecurrently_pregnant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;no or unsure\u0026apos;,\u0026apos;yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0,1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003enumber_children\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;\u0026lt;4\u0026apos;,\u0026apos;\u0026gt;=4\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003enumber of children \u0026gt; 3\u0026rsquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0,1\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003edesire_morechildren\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;Wants no more\u0026apos;, \u0026apos;Undecided\u0026apos;,\u0026apos;Wants\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003edesire_morechildren\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;Wants no more\u0026apos;, \u0026apos;Undecided\u0026apos;,\u0026apos;Wants\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0,1,2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003efather_education\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;no education\u0026apos;, \u0026apos;primary\u0026apos;, \u0026apos;secondary\u0026apos;,\u0026apos;higher\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003eFather_uneducated\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0,1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003efather_occupation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;Not working\u0026apos;,\u0026apos;Working\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026lsquo;father_unemployed\u0026rsquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003eMother_occupation\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;Not working\u0026apos;,\u0026apos;Working\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003eMother_ unemployed\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003eSex_child\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;male\u0026apos;,\u0026apos;female\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026lsquo;male-child-sex\u0026rsquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003eNo, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003echid_age\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;6-11\u0026apos;,\u0026apos;12-23\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003echid_age\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;6-11\u0026apos;,\u0026apos;12-23\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003eANC_visit\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003eANC_visit\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003ePlace_delivery\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;Home\u0026apos;,\u0026apos;Health facility\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026lsquo;home delivery\u0026rsquo;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003emode_delivery\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;Spontanous vaginal delivery\u0026apos;,\u0026apos;Cesarean section\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003e\u0026apos;cesarean section delivery\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003ePNC_visit\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;no\u0026apos;,\u0026apos;yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003ePNC_visit\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;no\u0026apos;,\u0026apos;yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003eHad_diarrhea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003eHad_diarrhea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003ehad_cough\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003ehad_cough\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003emedia_exposure\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003emedia_exposure\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 162px;\"\u003e\n \u003cp\u003eMicronutrient_intakestatus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 158px;\"\u003e\n \u003cp\u003eMicronutrient_intakestatus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026apos;No\u0026apos;,\u0026apos;Yes\u0026apos;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73px;\"\u003e\n \u003cp\u003e0, 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eAmong 2500 subjects, 503 had a cough, and the remaining 1997 subjects didn\u0026rsquo;t have one. This shows the data is highly imbalanced, and the results obtained and the models developed from it will be biased. So, synthetic minority oversampling techniques (SMOTE) were used to make more minority classes by interpolating between neighbors. This increased the number of instances of the minority class, which will help get rid of any possible bias in the machine learning model. The application of SMOTE led to an increase in the data to 3992, with 1996 elements in each class, including those with and without coughs. The tasks of evaluating features and building a machine learning model were done before and after the data balancing algorithm was used to see how the imbalance affected them.\u003c/p\u003e\n\u003cp\u003e3.2.\u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Feature assessment result \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe feature importances were computed using logistic regression and random forest algorithms. Table 2 shows the results obtained from feature importance calculations. After applying SMOTE, the variables mode of delivery, region, and current pregnancy status resulted in 0.09, 0.089, and 0.079 values, respectively.\u003c/p\u003e\n\u003cp\u003eTable 2: Feature importance results before and after applying SMOTE\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFeature\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 154px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBefore SMOTE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 154px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAfter SMOTE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eLR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eRF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eLR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eRF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003ematernal_age\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.069\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.039\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.062\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.035\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003erural-residence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eMother_uneducated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.036\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.022\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.031\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003efamily-size \u0026gt;5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.039\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.024\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.033\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.017\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003emale-househead\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.032\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.019\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.019\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003ewealth_index\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.056\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.047\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.025\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003ecurrently_pregnant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.045\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.029\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.026\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.022\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003enumber of children \u0026gt; 3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.016\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.031\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.025\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.015\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003edesire_morechildren\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.051\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.026\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.044\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.022\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eFather_uneducated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.015\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.027\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.034\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.021\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003efather_unemployed\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.028\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.024\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.017\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eMother_ unemployed\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.041\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.041\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.041\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.052\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.036\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003emale-child-sex\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.026\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.052\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.039\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.045\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003echid_age\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.006\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eANC_visit\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.034\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.033\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.022\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003ehome delivery\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.035\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.032\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003ecesarean section delivery\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.056\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.032\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.031\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.022\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003ePNC_visit\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.025\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.014\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.016\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.022\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.019\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eHad_diarrhea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.116\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.058\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.087\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.034\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.026\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003emedia_exposure\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.029\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.033\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.033\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eMicronutrient_intakestatus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.006\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.039\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.006\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_tigray\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.122\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.031\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.077\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.029\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.024\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_afar\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.035\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.024\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.046\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.032\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_amhara\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.025\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.014\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.053\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.037\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_oromia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.042\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.032\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.047\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.035\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_somali\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.059\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.034\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.073\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.031\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.052\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_benishangul\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.108\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.061\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.096\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.044\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.07\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_snnpr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.015\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.019\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.049\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.019\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.034\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_gambela\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.033\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.022\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.071\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.017\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.044\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_harari\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.066\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.082\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.027\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.054\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_addis adaba\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.045\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.027\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003eregion_dire dawa\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.056\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.014\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.035\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003ereligion_Muslim\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.006\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.017\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.034\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.026\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003ereligion_Orthodox\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.015\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.017\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.028\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003ereligion_Protestant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.006\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.016\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.017\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.027\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 139px;\"\u003e\n \u003cp\u003ereligion_Others\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.006\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.049\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e0.027\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eFigure 2 illustrates the SHAP bar plot before and after applying the SMOTE data balancing technique. Figure 3 demonstrates the most influential variables, and the variable region significantly influences the output.\u003c/p\u003e\n\u003cp\u003e3.2. Model development and testing results\u003c/p\u003e\n\u003cp\u003e1DCNN, RF, SVM, LR, GBC, KNN, and DT models were developed. The models were tested and got an AUC score of 0.81, 0.91, 0.77, 0.73, 0.72, 0.85, and 0.76, in that order, after the SMOTE data class balancing technique was used. Table 3 presents the performance results of the developed machine learning models before and after applying SMOTE. The performance results obtained before the application of the data class balancing technique were significantly lower than the results obtained after the application of data class balancing. This entails that data class balancing plays an essential role in overcoming high-level data class imbalance challenges while developing machine learning models. Based on the results obtained, the random forest model is the best performer with an AUC score of 0.91 and an accuracy of 0.83.\u003c/p\u003e\n\u003cp\u003eTable 3. The models\u0026rsquo; performance scores before and after applying data class balancing\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"492\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel name\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAccuracy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRecall\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eF1_score\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePrecision\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC-score\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\" valign=\"top\" style=\"width: 492px;\"\u003e\n \u003cp\u003eEvaluation results before applying SMOTE\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eLR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.799\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.114\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.186\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.508\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.700\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eSVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.797\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.034\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.062\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.504\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.641\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eKNN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.784\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.090\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.143\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.357\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.592\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eDT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.701\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.292\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.282\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.275\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.548\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eRF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.785\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.119\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.182\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.399\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.655\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eGBC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.798\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.692\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e1DCNN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.742\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.256\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.283\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.326\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.599\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\" valign=\"top\" style=\"width: 492px;\"\u003e\n \u003cp\u003eEvaluation results after applying SMOTE\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eLR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.758\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.745\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.754\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.764\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.842\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eSVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.804\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.790\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.801\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.813\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.881\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eKNN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.749\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.914\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.784\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.687\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.860\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eDT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.788\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.827\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.796\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.767\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.792\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eRF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.840\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.862\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.844\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.827\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.918\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eGBC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.668\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.716\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.683\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.658\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.726\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e1DCNN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e0.800\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 60px;\"\u003e\n \u003cp\u003e0.824\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.804\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.788\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e0.872\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eAs shown in Figure 4a, the ROC curve for the models using the data prior to SMOTE lies just above the straight line, meaning their performance is slightly better than random guessing but not very well. On the other hand, the ROC curves of the models developed using the balanced data given in Figure 4b showed improvements, and some of the models are performing very well. It is visible that the ROC curve of the random forest model is expanding to the left upper corner after applying data balancing.\u003c/p\u003e\n\u003cp\u003eFigure 5 is a confusion matrix plot that shows how many wrongly classified data points there were in models that were made with unbalanced data (Figure 5a) and balanced data (Figure 5b).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3. Deployment\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe machine learning backend application, developed in the form of web-based app (accessible by the link https://aridetector.streamlit.app/) and desktop app (can be accessed from https://drive.google.com/file/d/1iwuyaVyTbqYhh3bfOAGifLjc77Vp98TD/view) for predicting the probability of cough presence. The online app can be used directly by using the link, while the desktop version requires to download and install on the local computer. In both online and desktop applications, the key features for determining the possibility of cough presence based on demographic information were prepared for the users to select their corresponding weights or nominal values. Users can select the weightage or set nominal values for each factor in the applications. The applications also include a decision button that allows users to predict the likelihood of a cough by clicking the submit button. Finally, the system will display the prediction result. Figure 6(a) shows the online app outline, while Figure 6(b) shows the desktop app frontend.\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eARI is the leading cause of death and disability of children under the age of five years [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Malnutrition, inadequate immunization, poor sanitation, and limited access to healthcare facilities are among the main factors contributing to ARI in LMICs [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. ARIs, especially pneumonia, are the major healthcare challenges in Ethiopia [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. As traditional diagnostic and monitoring approaches are time-consuming and inaccessible to all early predictions, ML is becoming a more robust tool for early prediction [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Children 6 months to 2 years old are highly vulnerable to ARI, yet early prediction of ARI utilizing the robust ML algorithm is not explored [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. The aim of this study is to develop ML models for predicting the ARI in children aged 6 months to 2 years. The data obtained from the national repository contains 2500 entities; out of those, 503 had a cough, and the remaining in 1997 didn\u0026rsquo;t have a cough. The ratio of subjects having a cough to those who didn\u0026rsquo;t is almost 1:4, making the data highly imbalanced. Using this imbalanced data, LR, SVM, KNN, DT, RF, GBC, and 1DCNN models were developed and resulted in AUC scores of 0.700, 0.641, 0.592, 0.548, 0.655, 0.692, and 0.599 and recall scores of 0.114, 0.034, 0.090, 0.292, 0.119, 0.00, and 0.256, respectively. The results presented lack clarity. As Data class balance is crucial for developing machine learning models; therefore, SMOTE data balancing techniques were employed, resulting in 3992 data elements with 1996 entities in both the 'had-cough' and 'didn\u0026rsquo;t-have-cough' classes. Using the balanced data, AUC scores of 0.842, 0.881, 0.860, 0.792, 0.918, 0.726, and 0.872 and recall scores of 0.745, 0.790, 0.914, 0.827, 0.862, 0.716, and 0.824 were obtained for LR, SVM, KNN, DT, RF, GBC, and 1DCNN models, respectively.\u003c/p\u003e\u003cp\u003eUsing k-fold cross-validation, initially, before the data-balancing operation, the average number of data allocated for testing was 500. Out of 500 data elements, 100, 101, 118, 149, 107, 100, and 129 were wrongly classified by LR, SVM, KNN, DT, RF, GBC, and 1DCNN models, respectively. The percentages of wrongly classified cases were 20%, 20.2%, 23.6%, 29.8%, 21.4%, 20%, and 25.8%. Upon balancing the data classes, the testing set increased to 798 data elements, and 193, 156, 200, 169, 127, 264, and 159 were wrongly classified by LR, SVM, KNN, DT, RF, GBC, and 1DCNN models, respectively. The percentages of wrongly classified cases were 24%, 19.5%, 25%, 21.2%, 15.9%, 33%, and 19.9%. These results show a significant improvement in performance for random forests and support vector machines in classifying each entity correctly. The findings show that SMOTE data balancing techniques have great importance for overcoming the bias and alteration in the development of ML models. Additionally, performance results entail that from the demographic survey data, it is possible to predict the possibility of ARIs using ML models from balanced data.\u003c/p\u003e\u003cp\u003eA study done to develop a machine learning model from DHS data showed AUC scores of 0.79, 0.94, 0.73, 0.73, 0.90, 0.80, and 0.81 and sensitivities of 0.88, 0.87, 0.68, 0.72, 0.79, 0.68, and 0.77 for decision tree, random forest, Na\u0026iuml;ve Bayes, logistic regression, KNN, SVM, and gradient boosting models, respectively [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. With a recall of 0.87 and an AUC score of 0.94, the Forest algorithm performed best in predicting ARI from DHS data [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. A study showed a random forest ML model can predict ARI from DHS data with an accuracy of 0.96 [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. An ensemble model was able to predict the ARI with an accuracy of 0.86 and sensitivity of 0.84 [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. An AUC score of 0.95, a precision of 0.89, an accuracy of 0.88, an F-1 score of 0.83, and a recall of 0.77 were obtained from a random forest model trained with demographic health survey data [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Using the demographic health survey obtained in 2016 in the Ethiopian population, the XGBoost machine learning model was scoring an accuracy of 0.79 and an AUC score of 0.86 for predicting ARI [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. From these results, the RF model demonstrates a significant improvement in classification accuracy due to data-class balancing. Thus, the RF model was selected for GUI development. The developed GUI enables the users to select the nominal, ordinal, or range values for the given demographic health survey (DHS) data collection and predicts the percentage of a certain DHS entry to be classified as having a cough or not having a cough. Developing an online and desktop application from the best-performing machine learning model enables it to be evaluated thoroughly by the researchers, health care service providers, and regulatory authorities. These evaluations will make a great contribution to further improvement of the model and the GUI system. Then, finally, upon evaluation of the developed system by the regulatory authorities, this system can be used to predict the potential forthcoming cough from demographic health survey data.\u003c/p\u003e\u003cp\u003eResearchers were able to create and utilize machine learning applications for healthcare that work online and offline. These applications can predict if someone has COVID-19 based on user input [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e], manage electronic health records for long-term conditions [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e], visualize bioinformatics and cheminformatics projects [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e], predict cancer, heart diseases, diabetes, and kidney diseases using the Streamlit web hosting platform [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e], and estimate child deaths from pneumonia[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e].\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study utilized nationally representative Demographic and Health Survey (DHS) data to develop machine learning models for predicting acute respiratory infections (ARI) in children. Algorithms implemented included logistic regression, support vector machines, k-nearest neighbors, decision trees, random forests, gradient boosting classifiers, and one-dimensional convolutional neural networks (1D-CNN). Feature importance was assessed using random forest, logistic regression, and SHAP (Shapley Additive Explanations), identifying region, maternal unemployment, and paternal educational status as key predictors of ARI. Among all models, the random forest algorithm demonstrated superior performance based on accuracy, recall, and area under the curve (AUC) metrics. This best-performing model was subsequently deployed as a web application via Streamlit, with an offline version also prepared for Windows systems. These findings demonstrate the feasibility of leveraging demographic data and machine learning for early ARI prediction, particularly in resource-constrained settings.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eEthics approval and consent to participate\u003c/p\u003e\n\u003cp\u003eThis study is a secondary data analysis from the EDHS data, so it does not require ethical approval. For conducting this study, online registration and requests for measuring DHS were conducted. The dataset was downloaded from the DHS online archive from the MEASURE DHS dataset for free after getting approval to access the data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePatient and public involvement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study was conducted using secondary data and did not involve patient or public participation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no conflicts of interests\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors did not receive any funds for publication\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors’ contribution\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors have been involved in all steps from conceptualization, data extraction, analysis, and manuscript writing. All authors proofread the manuscript and approved its submission for publication.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eM. Cashat-Cruz, J. J. Morales-Aguirre, and M. Mendoza-Azpiri, \u0026ldquo;Respiratory tract infections in children in developing countries,\u0026rdquo; \u003cem\u003eSemin. \u003c/em\u003e\u003cem\u003ePediatr. Infect. Dis.\u003c/em\u003e, vol. 16, no. 2, pp. 84\u0026ndash;92, Apr. 2005, doi: 10.1053/j.spid.2005.12.005.\u003c/li\u003e\n\u003cli\u003e\u0026ldquo;Pneumonia in children.\u0026rdquo; Accessed: Mar. 12, 2025. [Online]. Available: https://www.who.int/news-room/fact-sheets/detail/pneumonia\u003c/li\u003e\n\u003cli\u003eJ. O. Akinyemi and O. M. Morakinyo, \u0026ldquo;Household environment and symptoms of childhood acute respiratory tract infections in Nigeria, 2003\u0026ndash;2013: a decade of progress and stagnation,\u0026rdquo; \u003cem\u003eBMC Infect. Dis.\u003c/em\u003e, vol. 18, no. 1, p. 296, July 2018, doi: 10.1186/s12879-018-3207-5.\u003c/li\u003e\n\u003cli\u003eS. Hassen \u003cem\u003eet al.\u003c/em\u003e, \u0026ldquo;Determinants of acute respiratory infection (ARI) among under-five children in rural areas of Legambo District, South Wollo Zone, Ethiopia: A matched case\u0026ndash;control study,\u0026rdquo; \u003cem\u003eInt. J. Infect. Dis.\u003c/em\u003e, vol. 96, pp. 688\u0026ndash;695, July 2020, doi: 10.1016/j.ijid.2020.05.012.\u003c/li\u003e\n\u003cli\u003eFaculty of Public Health, Universitas Indonesia and E. Triana, \u0026ldquo;Factors Affecting The Incidence of Acute Respiratory Tract Infection in Children under Five at Betungan Community Health Center, Bengkulu,\u0026rdquo; in \u003cem\u003eStrengthening Hospital Competitiveness to Improve Patient Satisfaction and Better Health Outcomes\u003c/em\u003e, Masters Program in Public Health, Graduate School, Universitas Sebelas Maret, 2019, pp. 40\u0026ndash;45. doi: 10.26911/the6thicph-FP.01.06.\u003c/li\u003e\n\u003cli\u003eM. Sultana \u003cem\u003eet al.\u003c/em\u003e, \u0026ldquo;Prevalence, determinants and health care-seeking behavior of childhood acute respiratory tract infections in Bangladesh,\u0026rdquo; \u003cem\u003ePLOS ONE\u003c/em\u003e, vol. 14, no. 1, p. e0210433, Jan. 2019, doi: 10.1371/journal.pone.0210433.\u003c/li\u003e\n\u003cli\u003eI. Mej\u0026iacute;a-Guevara, W. Zuo, E. Bendavid, N. Li, and S. Tuljapurkar, \u0026ldquo;Age distribution, trends, and forecasts of under-5 mortality in 31 sub-Saharan African countries: A modeling study,\u0026rdquo; \u003cem\u003ePLOS Med.\u003c/em\u003e, vol. 16, no. 3, p. e1002757, Mar. 2019, doi: 10.1371/journal.pmed.1002757.\u003c/li\u003e\n\u003cli\u003eS. Yaya, G. Bishwajit, F. Okonofua, and O. A. Uthman, \u0026ldquo;Under five mortality patterns and associated maternal risk factors in sub-Saharan Africa: A multi-country analysis,\u0026rdquo; \u003cem\u003ePLOS ONE\u003c/em\u003e, vol. 13, no. 10, p. e0205977, Oct. 2018, doi: 10.1371/journal.pone.0205977.\u003c/li\u003e\n\u003cli\u003eM. Sonego, M. C. Pellegrin, G. Becker, and M. Lazzerini, \u0026ldquo;Risk Factors for Mortality from Acute Lower Respiratory Infections (ALRI) in Children under Five Years of Age in Low and Middle-Income Countries: A Systematic Review and Meta-Analysis of Observational Studies,\u0026rdquo; \u003cem\u003ePLOS ONE\u003c/em\u003e, vol. 10, no. 1, p. e0116380, Jan. 2015, doi: 10.1371/journal.pone.0116380.\u003c/li\u003e\n\u003cli\u003eE. D\u0026iacute;az-Mart\u0026iacute;nez and E. D. and Gibbons, \u0026ldquo;The Questionable Power of the Millennium Development Goal to Reduce Child Mortality,\u0026rdquo; \u003cem\u003eJ. Hum. Dev. Capab.\u003c/em\u003e, vol. 15, no. 2\u0026ndash;3, pp. 203\u0026ndash;217, July 2014, doi: 10.1080/19452829.2013.864621.\u003c/li\u003e\n\u003cli\u003eA. Pr\u0026uuml;ss-Ust\u0026uuml;n \u003cem\u003eet al.\u003c/em\u003e, \u0026ldquo;Burden of disease from inadequate water, sanitation and hygiene for selected adverse health outcomes: An updated analysis with a focus on low- and middle-income countries,\u0026rdquo; \u003cem\u003eInt. J. Hyg. Environ. Health\u003c/em\u003e, vol. 222, no. 5, pp. 765\u0026ndash;777, June 2019, doi: 10.1016/j.ijheh.2019.05.004.\u003c/li\u003e\n\u003cli\u003eA. Keleb \u003cem\u003eet al.\u003c/em\u003e, \u0026ldquo;Pneumonia remains a leading public health problem among under-five children in peri-urban areas of north-eastern Ethiopia,\u0026rdquo; \u003cem\u003ePLOS ONE\u003c/em\u003e, vol. 15, no. 9, p. e0235818, Sept. 2020, doi: 10.1371/journal.pone.0235818.\u003c/li\u003e\n\u003cli\u003eT. Nigatu \u003cem\u003eet al.\u003c/em\u003e, \u0026ldquo;The status of immunization program and challenges in Ethiopia: A mixed method study,\u0026rdquo; \u003cem\u003eSAGE Open Med.\u003c/em\u003e, vol. 12, p. 20503121241237115, June 2024, doi: 10.1177/20503121241237115.\u003c/li\u003e\n\u003cli\u003eF. Epelde, \u0026ldquo;How AI Could Help Us in the Epidemiology and Diagnosis of Acute Respiratory Infections?,\u0026rdquo; \u003cem\u003ePathogens\u003c/em\u003e, vol. 13, no. 11, Art. no. 11, Nov. 2024, doi: 10.3390/pathogens13110940.\u003c/li\u003e\n\u003cli\u003eD. A. Rankin \u003cem\u003eet al.\u003c/em\u003e, \u0026ldquo;Diagnostic models predicting paediatric viral acute respiratory infections: a systematic review,\u0026rdquo; \u003cem\u003eBMJ Open\u003c/em\u003e, vol. 13, no. 4, p. e067878, Apr. 2023, doi: 10.1136/bmjopen-2022-067878.\u003c/li\u003e\n\u003cli\u003eP. Yadav, V. Rastogi, A. Yadav, and P. Parashar, \u0026ldquo;Artificial Intelligence: A promising tool in diagnosis of respiratory diseases,\u0026rdquo; \u003cem\u003eIntell. Pharm.\u003c/em\u003e, vol. 2, no. 6, pp. 784\u0026ndash;791, Dec. 2024, doi: 10.1016/j.ipha.2024.05.002.\u003c/li\u003e\n\u003cli\u003eS. Al-Anazi \u003cem\u003eet al.\u003c/em\u003e, \u0026ldquo;Artificial intelligence in respiratory care: Current scenario and future perspective,\u0026rdquo; \u003cem\u003eAnn. Thorac. Med.\u003c/em\u003e, vol. 19, no. 2, p. 117, June 2024, doi: 10.4103/atm.atm_192_23.\u003c/li\u003e\n\u003cli\u003eA. K. Kassaw, G. Bekele, A. K. Kassaw, and A. Yimer, \u0026ldquo;Prediction of acute respiratory infections using machine learning techniques in Amhara Region, Ethiopia,\u0026rdquo; \u003cem\u003eSci. Rep.\u003c/em\u003e, vol. 14, no. 1, p. 27968, Nov. 2024, doi: 10.1038/s41598-024-76847-3.\u003c/li\u003e\n\u003cli\u003eH. M. Fenta, T. T. Zewotir, S. Naidoo, R. N. Naidoo, and H. Mwambi, \u0026ldquo;Factors of acute respiratory infection among under-five children across sub-Saharan African countries using machine learning approaches,\u0026rdquo; \u003cem\u003eSci. Rep.\u003c/em\u003e, vol. 14, no. 1, p. 15801, July 2024, doi: 10.1038/s41598-024-65620-1.\u003c/li\u003e\n\u003cli\u003eY. Ku, S. B. Kwon, J.-H. Yoon, S.-K. Mun, and M. Chang, \u0026ldquo;Machine Learning Models for Predicting the Occurrence of Respiratory Diseases Using Climatic and Air-Pollution Factors,\u0026rdquo; \u003cem\u003eClin. Exp. Otorhinolaryngol.\u003c/em\u003e, vol. 15, no. 2, pp. 168\u0026ndash;176, Jan. 2022, doi: 10.21053/ceo.2021.01536.\u003c/li\u003e\n\u003cli\u003eR. M. Kananura, \u0026ldquo;Machine learning predictive modelling for identification of predictors of acute respiratory infection and diarrhoea in Uganda\u0026rsquo;s rural and urban settings,\u0026rdquo; \u003cem\u003ePLOS Glob. Public Health\u003c/em\u003e, vol. 2, no. 5, p. e0000430, May 2022, doi: 10.1371/journal.pgph.0000430.\u003c/li\u003e\n\u003cli\u003eT. Z. Yehuala, B. M. Fente, S. M. Wubante, and N. M. Derseh, \u0026ldquo;Exploring machine learning algorithms to predict acute respiratory tract infection and identify its determinants among children under five in Sub-Saharan Africa,\u0026rdquo; \u003cem\u003eFront. Pediatr.\u003c/em\u003e, vol. 12, Nov. 2024, doi: 10.3389/fped.2024.1388820.\u003c/li\u003e\n\u003cli\u003eM. H. Kalayou, A.-A. K. Kassaw, and K. B. Shiferaw, \u0026ldquo;Empowering child health: Harnessing machine learning to predict acute respiratory infections in Ethiopian under-fives using demographic and health survey insights,\u0026rdquo; \u003cem\u003eBMC Infect. Dis.\u003c/em\u003e, vol. 24, no. 1, p. 338, Mar. 2024, doi: 10.1186/s12879-024-09195-2.\u003c/li\u003e\n\u003cli\u003eT.-H. Chang \u003cem\u003eet al.\u003c/em\u003e, \u0026ldquo;Clinical characteristics of hospitalized children with community-acquired pneumonia and respiratory infections: Using machine learning approaches to support pathogen prediction at admission,\u0026rdquo; \u003cem\u003eJ. Microbiol. Immunol. Infect.\u003c/em\u003e, vol. 56, no. 4, pp. 772\u0026ndash;781, Aug. 2023, doi: 10.1016/j.jmii.2023.04.011.\u003c/li\u003e\n\u003cli\u003eZ. B. Tadese \u003cem\u003eet al.\u003c/em\u003e, \u0026ldquo;Interpretable prediction of acute respiratory infection disease among under-five children in Ethiopia using ensemble machine learning and Shapley additive explanations (SHAP),\u0026rdquo; \u003cem\u003eDigit. Health\u003c/em\u003e, vol. 10, p. 20552076241272739, Sept. 2024, doi: 10.1177/20552076241272739.\u003c/li\u003e\n\u003cli\u003e\u0026ldquo;Leveraging Machine Learning for Predictive Models in Healthcare to Enhance Patient Outcome Management,\u0026rdquo; \u003cem\u003eInt. Res. J. Mod. Eng. Technol. Sci.\u003c/em\u003e, Jan. 2025, doi: 10.56726/IRJMETS66198.\u003c/li\u003e\n\u003cli\u003eA. Rahman, M. Karmakar, and P. Debnath, \u0026ldquo;Predictive Analytics for Healthcare: Improving Patient Outcomes in the U.S. through Machine Learning,\u0026rdquo; \u003cem\u003eRev. Intel. Artif. En Med.\u003c/em\u003e, vol. 14, no. 1, Art. no. 1, Nov. 2023.\u003c/li\u003e\n\u003cli\u003eH. Ali, \u0026ldquo;AI for Pandemic Preparedness and Infectious Disease Surveillance: Predicting Outbreaks, Modeling Transmission, and Optimizing Public Health Interventions,\u0026rdquo; \u003cem\u003eInt. J. Res. Publ. Rev.\u003c/em\u003e, vol. 6, no. 2, pp. 4605\u0026ndash;4619, Feb. 2025, doi: 10.55248/gengpi.6.0225.0941.\u003c/li\u003e\n\u003cli\u003eT. Z. Yehuala \u003cem\u003eet al.\u003c/em\u003e, \u0026ldquo;Machine learning algorithms to predict healthcare-seeking behaviors of mothers for acute respiratory infections and their determinants among children under five in sub-Saharan Africa,\u0026rdquo; \u003cem\u003eFront. Public Health\u003c/em\u003e, vol. 12, June 2024, doi: 10.3389/fpubh.2024.1362392.\u003c/li\u003e\n\u003cli\u003eC. N. Villavicencio, J. J. Macrohon, X. A. Inbaraj, J.-H. Jeng, and J.-G. Hsieh, \u0026ldquo;Development of a Machine Learning Based Web Application for Early Diagnosis of COVID-19 Based on Symptoms,\u0026rdquo; \u003cem\u003eDiagnostics\u003c/em\u003e, vol. 12, no. 4, Art. no. 4, Apr. 2022, doi: 10.3390/diagnostics12040821.\u003c/li\u003e\n\u003cli\u003eS. Burns, A. Cushing, A. Taylor, D. J. Lowe, and C. Carlin, \u0026ldquo;Supporting long-term condition management: a workflow framework for the co-development and operationalization of machine learning models using electronic health record data insights,\u0026rdquo; \u003cem\u003eFront. Artif. Intell.\u003c/em\u003e, vol. 7, p. 1458508, Nov. 2024, doi: 10.3389/frai.2024.1458508.\u003c/li\u003e\n\u003cli\u003eC. Nantasenamat, A. Biswas, J. M. Napoles, M. Parker, and R. Dunbrack, \u0026ldquo;Building bioinformatics web applications with Streamlit,\u0026rdquo; 2023, pp. 679\u0026ndash;699. doi: 10.1016/B978-0-443-18638-7.00001-3.\u003c/li\u003e\n\u003cli\u003eL. D. Gopisetti, S. K. L. Kummera, S. R. Pattamsetti, S. Kuna, N. Parsi, and H. P. Kodali, \u0026ldquo;Multiple Disease Prediction System using Machine Learning and Streamlit,\u0026rdquo; in \u003cem\u003e2023 5th International Conference on Smart Systems and Inventive Technology (ICSSIT)\u003c/em\u003e, Jan. 2023, pp. 923\u0026ndash;931. doi: 10.1109/ICSSIT55814.2023.10060903.\u003c/li\u003e\n\u003cli\u003eN. I. Mohammed, A. Jarde, G. Mackenzie, U. D\u0026rsquo;Alessandro, and D. Jeffries, \u0026ldquo;Deploying Machine Learning Models Using Progressive Web Applications: Implementation Using a Neural Network Prediction Model for Pneumonia Related Child Mortality in The Gambia,\u0026rdquo; \u003cem\u003eFront. Public Health\u003c/em\u003e, vol. 9, p. 772620, Feb. 2022, doi: 10.3389/fpubh.2021.772620.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-informatics-and-decision-making","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"midm","sideBox":"Learn more about [BMC Medical Informatics and Decision Making](http://bmcmedinformdecismak.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/midm/default.aspx","title":"BMC Medical Informatics and Decision Making","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7425776/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7425776/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eIntroduction\u003c/strong\u003e: Acute respiratory infections (ARI) caused by various pathogens are the cause of millions of illnesses and deaths among children under five. The prevalence of ARI is higher in low- and middle-income countries. To this date, in low- and middle-income countries, the management of ARI in children under the age of two is mainly curative, not preventive. Thus, this study aimed to explore the capability of machine learning models to predict the forthcoming ARI from the general demographic health survey data by developing and deploying predictive machine learning models.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e: The demographic health survey data was obtained from the USAID repository, the data was preprocessed, and the important features were identified. Then data class balancing was done using synthetic minority oversampling techniques. Then, logistic regression, support vector machine, k-nearest neighbor, decision tree, random forest, gradient boosting, and one-dimensional convolutional neural network models were developed. The K-fold cross-validation technique was used to train the model and obtain a stable model and representative performance metrics. The accuracy, the recall, the F1 score, the precision, and the AUC score results were calculated and used to select the best-performing model. Finally, the selected model was deployed on Streamlit as a web-based application and using the Python tkinter library for developing desktop applications.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e: A total of 2500 subjects’ data were obtained, out of which 503 subjects were having coughs, which is nearly one-fifth of the total data. Upon applying the synthetic minority oversampling technique (SMOTE), the overall data is increased to 3992, with each class having 1996 subjects’ data. At first, the data had 23 features, but after changing some features from categories to numbers and giving numerical values to ordered and yes/no features, there were 36 features in total. Following data class balancing and data preprocessing, seven models were trained and resulted in AUC scores of 0.842, 0.881, 0.860, 0.792, 0.918, 0.918, 0.918, 0.726, and 0.872, and recall scores of 0.745, 0.790, 0.914, 0.827, 0.862, 0.716, and 0.824 were obtained for LR, SVM, KNN, DT, RF, GBC, and 1DCNN models, respectively. Then the best-performing model, which is the random forest model, was selected and deployed as a web-based application on Streamlit and as an offline Windows application using the Python tkinter library.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion\u003c/strong\u003e: This study illustrates the possibilities of machine learning backend applications for predicting the forthcoming ARI from the demographic health survey data, which will play a key role in preventing diseases upon necessary regulatory and quality checks. In low-resource setting areas that are highly vulnerable to ARI, machine learning-based applications will be useful. Further studies need to be done considering a wider range of parameters for improving the predictability and accuracy of the models.\u003c/p\u003e","manuscriptTitle":"Acute respiratory infections risk prediction using machine learning among Ethiopian children Aged 6 Months to 2 Years","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-09 12:33:58","doi":"10.21203/rs.3.rs-7425776/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"257038591497765660864904277584386491953","date":"2025-10-04T15:42:50+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"171257081199747048912503422436744346647","date":"2025-09-27T08:11:56+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-26T21:07:42+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"307630224331933861995328310932684898419","date":"2025-09-25T08:27:43+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-18T08:31:13+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-08-25T10:34:26+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-22T12:07:29+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-22T12:06:01+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Informatics and Decision Making","date":"2025-08-21T11:27:45+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-informatics-and-decision-making","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"midm","sideBox":"Learn more about [BMC Medical Informatics and Decision Making](http://bmcmedinformdecismak.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/midm/default.aspx","title":"BMC Medical Informatics and Decision Making","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3a7b6e4f-e427-4668-91fa-ae238d1bdd6f","owner":[],"postedDate":"December 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-12-09T12:33:58+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-09 12:33:58","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7425776","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7425776","identity":"rs-7425776","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}