Explainable machine learning algorithm to identify predictors of intention to use family planning among reproductive-age women in Ethiopia: Evidence from the performance monitoring and accountability (PMA) survey 2021 dataset

Explainable machine learning algorithm to identify predictors of intention to use family planning among reproductive-age women in Ethiopia: Evidence from the performance monitoring and accountability (PMA) survey 2021 dataset | 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 Explainable machine learning algorithm to identify predictors of intention to use family planning among reproductive-age women in Ethiopia: Evidence from the performance monitoring and accountability (PMA) survey 2021 dataset Jibril Bashir Adem, Tewodros Desalegn Nebi, Agmasie Damtew Walle, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3848375/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Introduction: Approximately 225 million people in developing nations wish to delay or cease childbearing, but do not use any form of contraception. In the least developed countries, contraceptive usage was significantly lower, at 40%, and was particularly low in Africa at 33%. It is widely believed that intentions are a strong predictor of behavior, and many interventions that aim to change behavior including that targeting family planning use rely on evaluating program effectiveness through analyzing behavioral intentions. Understanding a woman's intention to use contraceptive methods is crucial in predicting and promoting the use of such methods. Therefore, this study aims to assess the determinants of intention to use family planning among reproductive age women in Ethiopia using explainable machine learning algorithm Method Secondary data from the cross-sectional household and female survey conducted by PMA Ethiopia in 2021 were used in the study. Using Python 3.10 version software, eight machine learning classifiers were used to predict and identify significant determinants of intention to use family planning on a weighted sample of 5993 women. Performance metrics were used to evaluate the classifiers. To smooth the data for additional analysis, data preparation techniques such as feature engineering, data splitting, handling missing values, addressing imbalanced categories, and outlier removal were used. Lastly, the greatest predictors of intention to utilize family planning were found using Shapley Additive exPlanations (SHAP) analysis, which further clarified the predictors' impact on the model's results. Result Using tenfold cross-validation and balanced training data, Random Forest revealed a performance of 77.0% accuracy and 85% areas under the curve, making it the most effective prediction model. The age at which family planning was first used, a partner or husband older than 40, being single, being Muslim, being pregnant, having previously been pregnant, needing to have more children, having a son or daughter relationship to the head of the household, and unmet needs for spacing and limiting were the top predictors of intention to use family planning, according to the SHAP analysis based on the random forest model. The research findings indicate that a range of personal and cultural factors may be taken into account when enacting health policies to enhance family planning intentions in Ethiopia. Therefore it’s highly recommended that the intention of family planning use and initiation of family planning provision should become a standard of service delivery to achieve the 2030 SDGs. Machine learning algorithm Intention to use family planning Predictors Ethiopia Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction In the next three decades, the population of the world is expected to increase by almost 40% ( 1 ). Over half of the additional 3 billion people will be in developing regions like sub-Saharan Africa and Asia, where the majority of the population growth is anticipated ( 2 ). Out of the 1.9 billion women of reproductive age in the globe, 1.1 billion need family planning services, and 190 million women wish to prevent getting pregnant but do not use any form of contraception( 3 , 4 ). Overall, there is a considerable variation in terms of contraceptive prevalence rate among different methods, ranging from 4.6% in South Sudan to 72.1% in Canada ( 3 ). In 2020, approximately 225 million people in developing nations wish to delay or cease childbearing, but do not use any form of contraception ( 5 ). In the least developed countries, contraceptive usage was significantly lower, at 40%, and was particularly low in Africa at 33% ( 6 ). The utilization of family planning is a significant predictor of preventing unwanted pregnancies, reducing maternal and child mortality, and enhancing the wellbeing of women and their families ( 7 ). Moreover, it is critical to achieving the 2030 Sustainable Development Goals (SDGs) ( 8 ). Globally, the average total fertility rate varies from 1.7 children per woman in the most developed countries to 4.6 in the least developed countries ( 9 ) Within Sub-Saharan Africa (SSA), total fertility rates range from 2.9 in Botswana to 7.2 in Niger ( 10 , 11 ). The high unmet need for contraception across all age groups and marital statuses accounts for both the rapid population growth and the steady increase in fertility ( 10 ). With an estimated 120 million people and an average annual growth rate of 2.6%, Ethiopia is the second most populous country in Sub-Saharan Africa ( 12 ). By 2050, Ethiopia's population is expected to reach 166 million, making it the tenth most populous nation in the world ( 9 ). Having a high overall fertility rate, high rates of maternal and infant mortality, and low rates of contraceptive use, Ethiopia is the second most populous country in Sub-Saharan Africa ( 12 ). Total fertility rate in Ethiopia is 5.3 children per woman ( 9 ). As a result, Ethiopia is now among the nations with the highest global fertility rates. Low family planning utilization is a major contributing factor to the high levels of fertility, particularly in less developed nations like Ethiopia ( 13 ). The proportion of married women in Ethiopia who use contraception ranges from 5% in the Somalia region to 53% in Addis Ababa and the Amhara region. The overall modern contraceptive prevalence rate in the country is 40% ( 4 ). Intention to use and use of family planning has been proven to be an effective method of controlling family size and reducing unintended pregnancies( 14 ). It is widely believed that intentions are a strong predictor of behavior, and many interventions that aim to change behavior including those targeting contraceptive use rely on evaluating program effectiveness through analyzing behavioral intentions ( 15 ). Understanding a woman's intention to use contraceptive methods is crucial in predicting and promoting the use of such methods ( 16 ). By doing so, we can improve the health of not just women but also children, families, and even societies ( 17 ). Numerous studies conducted on the use and intention to use family planning among reproductive age women have revealed factors such as switching between different forms of contraception ( 18 ), poor support from husband( 19 ), maternal age ( 18 ), maternal education ( 4 , 20 ), positive attitude to contraceptive use, occupation, knowledge of contraceptives ( 19 , 21 ), discussion on family planning with husband( 18 ), myths and misconceptions regarding contraception( 20 ), time of birth interval( 4 ), joint fertility decision( 22 ), and desire for live children( 19 ) among the variables influencing women of reproductive age's intention to use family planning. Similarly, the intention of family planning use among reproductive age women was significantly associated with, number of live children, and counseling during antenatal care, husbands' approval of family planning use, and having good knowledge of postpartum family planning( 23 , 24 ). Traditional regression models were previously used in Ethiopia to demonstrate the effects of socioeconomic, demographic, behavioral, maternal, and service-related characteristics on the intention of reproductive-age women to use family planning; however, the validity of these results declined with an increase in the number of variables and potential correlations ( 25 – 29 ). The multidisciplinary relationships between variables and numerous factors are typically problematic for these traditional models ( 30 , 31 ). Hence, in contrast to those traditional models, machine learning (ML) provides an effective way to find relevant characteristics linked to specific health outcomes for conducting public health research ( 30 , 32 , 33 ). Therefore, the purpose of this study is to evaluate the factors influencing Ethiopian women of reproductive age's intention to use family planning. This study seeks to determine and uncover consistent variables as well as new factors that influence the intention to utilize family planning using data from Ethiopian women of reproductive age in the PMA Survey 2021 dataset. The Ethiopian Ministry of Health and other health partners will be able to enhance the intention of reproductive-age women in Ethiopia to utilize family planning by concentrating on the most consistent and influential variables that will be designated as priority areas for intervention. Method Study design A machine learning (ML) technique was applied using secondary data from the PMA Ethiopia 2021 cross-sectional household and female Survey. PMA-Ethiopia is a five-year (2019–2023) project in collaboration with Addis Ababa University, Johns Hopkins University, and the Federal Ministry of Health. It consists of three distinct study activities: yearly cross-sectional surveys of women aged 15–49; longitudinal surveys of women who are pregnant or have given birth; and yearly service delivery point surveys of health facilities ( 34 ). Source and study population All 15–49 aged women in Ethiopia. Sample size The sample size in the present study was weighted for taking non-response and variations in the probability of selection into consideration. Further, only women who were of reproductive age at the time of the survey were included in the sample. Therefore, in a weighted sample, only 5993 women of reproductive age were eligible to be included in the analysis. Study variables Dependent variable The dependent variable was the intention to use family planning, which was divided into two groups: "intended to use" and "not intended to use." Predictor variables Sociodemograpic and economic factors including residence, women's age, region, and religion, size of family, education level, financial status, and media access are predictive of the intention to use family planning. The following reproductive health and family planning service characteristics were also included as predictors of intention to use family planning: partner or husband's feelings toward family planning, knowledge of all available contraceptive methods, age at first sex, having ever been pregnant, having ever used family planning (FP) methods, having ever been delivered in a health facility (HF), and partner being told not to use FP. Data processing and analysis In order to predict the intention to use family planning, this study employed the general framework found in previous research, which is based on Yufeng Guo's 7 Steps of Machine Learning. The following seven steps in supervised machine learning are outlined in the framework: data preparation, model training, model evaluation, parameter tuning, model selection, data collection, and prediction ( 35 ). Machine learning (ML) algorithms were implemented in Python 3.10.2 using Jupyter Notebook through Scikit-learn, and XGBoost packages. Data source/collection The dataset for this study is available on the PMA Survey website and can be obtained upon a formal request. A weighted sample of 5993 reproductive-age women is included in the study. Data preparation/pre-processing This study employed a variety of data preparation techniques, including feature engineering, data splitting, and data cleaning. Missing values were imputed using the R package 'CALIBERrfimpute' using Multivariate Imputation by Chained Equations (MICE) and random forest and the training data was balanced using the Synthetic Minority Oversampling Technique (SMOTE)( 36 ). SMOTE creates synthetic observations by interpolating between minority classes samples in the feature space. The Kolmogorov-Smirnov test indicates that the synthetic and existing observations differ significantly from each other (KS test statistic of 0.8214, p-value = < 0.001). Using Pandas get-dummies, One-Hot-Encoding techniques were applied to encode categorical data into dummy variables, and the relationship between the predictors and the outcome variable was evaluated using the SHAP feature importance method. A tenfold cross-validation technique was used to train the models( 36 ). Model selection and training We used eight different machine learning algorithms namely AdaBoost, logistic regression (LR), random forest (RF), K-nearest neighbor (KNN), artificial neural network, support vector machine (SVM), Naïve Bayes, and extreme gradient boosting (XGBoost) to predict intention to use family planning among reproductive age women and then determined which model was best suited for the problem( 36 ). Tenfold cross-validation was used to compare the performance of the chosen classifiers, and the best predictive model was chosen after comparison, adjusted with its hyper-parameters, and trained using balanced training data. Model evaluation and Hyper-parameter tuning Following training, the model was evaluated using the area under the receiver operating characteristic curve (AUC) score and classification accuracy. Moreover, ML model performance was shown using the receiver operating characteristic (ROC) curve. The Optuna framework was used to adjust the best model's hyper-parameters. It uses the Bayesian framework to better understand the probability of the optimal values and prevents needless computation for the combination of non-performing parameters in the search( 36 ). Prediction Prediction involves estimating the result variable using pre-selected predictors and applying the fully trained model to its intended output. In this study, the intention to utilize family planning was ascertained using the top predictors identified during the analysis. Consequently, the most accurate classifier will determine, to a predetermined degree of accuracy, whether or not a woman intends to use family planning. Model interpretation/explanation using Shapley Additive exPlanations (SHAP) An interpretation and explanation of any machine learning model's prediction can be found globally or locally using Shapley Additive exPlanations (SHAP) analysis. The fundamental concept behind the SHAP analysis is figuring out each predictor's marginal contribution to the outcome variable prediction result. In machine learning, features are chosen using SHAP values. The aggregate Shapley value of each characteristic for each sample should be shown in order to assess the impact of each predictor on the prediction of intention to use family planning. Moreover, a waterfall plot was used to illustrate how each component contributed to the prediction of a positive class. Finally, the overall data preparation and analysis process is presented in (Fig. 1 ). A thorough explanation of the SHAP analysis can be found in the previous article ( 36 ) Results Socio-demographic and economic characteristics The study included a weighted sample of 5993 women who were of reproductive age. The mean age of the sample was 23.89 years (± 19.2 standard deviation), with the majority of the women, 5,898 (71.87%), ranging between the ages of 15 and 19. 5120 (63.42%) of the participants were married or living with their spouses, while 4,957 (60.41%) were rural residents. Muslims made up 31.91% of the attendees, while Orthodox Christians made up the majority with 3,068 (38.00%). The majority of women, 2,595 (32.14%), have completed secondary school or above (Table 1 ). Table 1 socio-demographic status of reproductive age women in Ethiopia, 2021 Variable Categories Weighted Freq. Percent (%) Residence Urban 3,249 39.59 Rural 4,957 60.41 Marital status Married 5,120 63.42 single 2,220 27.50 Divorced/widowed 733 9.08 Women’s age 15 to 19 5,898 71.87 20 to 29 1,317 16.05 30 to 39 712 8.68 40 to 49 279 3.40 Education level No education 2,421 29.99 Primary 3,057 37.87 secondary and above 2,595 32.14 Wealth status Poor 2,742 33.42 Middle 1,386 16.89 Rich 4,077 49.69 Media access No 5,458 67.78 Yes 2,594 32.22 Religion Orthodox 3,068 38.00 Muslim 2,576 31.91 Protestant 2,363 29.27 Others* 66 0.82 Note: others* indicates: catholic, Wakefata, atheist and traditional Regarding the regional distribution of respondents, The Amhara region 1,584(19.30%), the Southern Nations, Nationalities, and Peoples' region 1,322 (16.11%), and the Oromia region accounted for the majority of respondents (1,858 (22.64%). No individuals from the Tigray regional state were in the study due to the conflict in the area throughout the study period (Fig. 2 ). Reproductive health and Family planning service characteristics Of the total respondents, 5,381 (60.41%) had previously been pregnant, and about 11.82% of the women (49.13%) had unmet family planning needs. About (48.3%) of the respondents had never utilized any family planning techniques. approximately 4,540 (74.69%) of the women started experiencing sex between the ages of 15 and 19, and 5.85% of the individuals said that their partner forced them into getting pregnant (Table 2 ). Table 2 Reproductive health and family planning service characteristics of study participants Variable Categories Weighted Freq. Percent (%) Ever been pregnant No 2,691 33.34 Yes 5,381 60.41 Ever used family planning No 3,876 48.31 Yes 4,148 51.69 Visited a facility for care for self or children in past 12 months No 4,184 51.96 yes 3,868 48.04 Current used of any contraceptive methods No 5,933 74.27 Yes 2,055 25.73 unmet need for family planning No 7,116 88.18 Yes 954 11.82 Your partner tried to force you to become pregnant No 4,815 94.15 Yes 299 5.85 Was visited by HEW who talked about FP in past 12 months No 7,370 91.59 Yes 677 8.41 Age at first sexual intercourse 15 to 19 4,539 74.69 20 to 29 1,478 24.32 30 to 39 55 0.91 40 to 49 5 0.08 Intention to use family planning among reproductive age women in Ethiopia, based on PMA survey, 2021 Among the study participants, 2,891(48.24%) reproductive age women had no intention of using any form of family planning (Fig. 3 ). Machine learning approach to identify predictors of intention to use family planning among reproductive age women in Ethiopia, 2021 PMA survey Balancing data and Model performance comparison In order to balance the outcome variable's unequal distribution, 168 additional synthetic observations from the minority category those who did not intend to utilize family planning were produced by the SMOTE oversampling approach. In order to create symmetric distributions for both categories and enable the development of reliable prediction models, the total intention to use status distribution was adjusted from 2313 not intended to use and 2481 intended to use, resulting in 2481 in each class. To assess the performance of the models to predict intention family planning among reproductive age women, a stratified 10-fold cross-validation was employed, with particular consideration paid to mean accuracy and mean area under the curve score. Stratified 10-fold cross-validation was used to assess classifiers on the unbalanced training data, and the support vector machine emerged to the top with an accuracy of 76.7% and 84.2% areas under the ROC curve. The unbalanced class character of the outcome variable, which could skew the model in favor of the majority class, makes this finding potentially unreliable. After balancing the training data using the SMOTE oversampling technique, an ML model comparison was conducted in order to prevent this biased model creation. Accordingly, with a 76.8% accuracy rate and an 84.6% area under the ROC curve, Random Forest emerged as the most accurate predictive model to predict intention to use family planning among reproductive age women in Ethiopia (Table 3 ). Table 3 Model comparison through tenfold cross-validation on training data Models Performance Unbalanced data Balanced data Logistic regression Accuracy (%) 76.3 75.7 AUC (%) 83.1 833 SVM Accuracy (%) 76.7* 76.7 AUC (%) 84.2* 84.3 XGBoost Accuracy (%) 75.3 76.3 AUC (%) 82.9 83.8 KNN Accuracy (%) 72.5 72.2 AUC (%) 78.6 78.7 Random forest Accuracy (%) 76.1 76.8* AUC (%) 83.5 84.6* AdaBoost Accuracy (%) 75.7 75.9 AUC (%) 82.4 83.2 Naïve Bayes Accuracy (%) 72.5 71.7 AUC (%) 77.3 77.2 Artificial neural network Accuracy (%) 72.6 73.5 AUC (%) 79.6 80.7 Hyper-parameter tuning of Random Forest Scikit-learn do not always provide the best solution for a given problem, even if it offers a set of reasonable default hyper-parameters for all models, including Random Forest. The number of decision trees in the forest (n_estimators), the number of features that each tree considers when splitting a node (max_features), the minimum number of samples needed to split an internal node (min_samples_split), the minimum number of samples needed to be at a leaf node (min_samples_leaf), and the number of samples to draw from independent variables to train each tree (max_samples) were among the hyper-parameters that were optimized with one hundred seventy five trials on a specified search space using stratified 10-fold cross-validation in order to maximize the performance of random forests. The Scikit-learn default hyper-parameters and our adjusted hyper-parameters are displayed in (Table 4 ). Table 4 Default and optimal tuned hyper-parameters of Random Forest model Hyper-parameters Default Optimal value Number of tree 100 175 Number of features considered for the best split Square root of the number of features 0.302 minimum number of samples required to split an internal node 2 9 minimum number of samples required to be at a leaf node 1 1 number of samples to draw from independent variables to train each base estimator None 1 Finally, random forest model was created with these tuned hyper-parameters on balanced training data through 10-fold cross-validation and yielded 77.0% accuracy and 85% areas under the curve. Feature selection Following the selection of the best model, previously unseen test data was used to predict the intention to utilize family planning. Following random forest training on unbalanced training data, balanced data using default model parameters, and a comparison with an optimized model trained on balanced data, the prediction was made. Following balanced and unbalanced data training for the random forest model, the prediction on unseen test data yielded an area under the ROC curve score of 0.83 for both. An improved AUC of 0.85 was anticipated by hyper-parameter-tuned random forest, nevertheless. In this study, the top predictors of the intention to use family planning were determined using model-agnostic SHAP global feature importance. In order to quantify a feature's contribution to the anticipated intention to use family planning among reproductive age women in Ethiopia, this technique looks at the mean absolute SHAP value for each predictor over all of the data. Higher mean absolute SHAP values indicate greater influence of the predictors, which are arranged in descending order of their impact on the outcome variable prediction. The findings showed that the most significant predictors of women's intention to use family planning were ever been utilized family planning (fp_ever_used_1), a partner or husband older than 40 (partner age_2), being single (marital_status_5), and being Muslim (religion_4). Other significant predictors of intention to use family planning included being pregnant (pregnant_1), having ever been pregnant (ever_pregnant_1), need to have more children (more_children_pregnant_2), having a son or daughter relationship to the head of the household (relationship_3), and unmet need for spacing and limiting. Each class's horizontal rectangle is half-filled with the colors red and blue, as it’s shown in the Fig. 6. As a result, each feature has an equal influence on whether a case is classified as intended to use family planning (label = 1) or not intended to use family planning (label = 0) (Fig. 4). Model interpretation/explanation A comprehensive picture of how the factors affect the model's predictions across the board was shown by using Beeswarms plots. Plotting the Shapley value of each individual predictor for each sample, Fig. 7 shows the distribution of the impacts of each predictor on the model's output (i.e., intention to use family planning prediction). The points on this beeswarm plot represent Shapley values of the features related to intention to use family planning status, providing insight into the importance and association of each of the top ten features on the outcome variable. The red and blue hues in the figure represent the higher and lower values of each predictor’s variable. Points that are right to the vertical line (0 SHAP value) increase the likelihood of intention to use family planning while the left side decreases likelihood of intention to use family. Since all of the variables are categorical and have two categories, the blue line denotes category code 0 (low value) and the red line denotes category code 1, which is a high value. Therefore, having a partner or spouse who is older than 40 (partner_Age_2), being a Muslim (Religion_4), having ever been pregnant (ever_pregnant_1), and being older than 30 while using family planning for the first time will increase the likelihood of intending to use family planning. Contrarily, the likelihood of intending to use family planning decreases with the following factors: having ever used family planning (fp_ever_used_1), being single (marital_status_5), having a partner between the ages of 21 and 39 (partner_age_1), being pregnant (pregnant_1), need to have more children (more_children_pregnant_2), and no unmet need for spacing and limiting (Unmate_need_for_spacing_limiting_1) (Fig. 5 ). Discussion Tenfold cross-validation was used to train eight machine learning classifiers on balanced and imbalanced training data. Random Forest was found to be the most accurate predictive model, and it was optimized for its optimal hyper-parameters before more analysis was carried out. Based on the result of this study, 51% of reproductive age women in Ethiopia had intention of using any form of family planning. This finding is higher than the result of a study conducted on Ethiopian Demographic and Health Survey, 2011 (44.1%) ( 37 ) but lower than the result of studies conducted in 90%( 38 ), Gahanna ( 39 ) and Sodo( 40 ) the prevalence was found to be 70%, Addis Ababa 60% ( 41 ). The reason behind the lower magnitude of intention to use contraception was the difference in availability, accessibility, infrastructure, and socioeconomic status between the present study, and a study done in America. The SHAP analysis revealed that age at first family planning use, a partner or husband older than 40, being single, being Muslim, being pregnant, having previously been pregnant, need to have more children, having a son or daughter relationship to the head of the household and need for spacing and limiting were significant predictors. According to the SHAP model explanation being older than 30 while using family planning for the first time will increase the likelihood of intending to use family planning. This finding is in line with the finding of study conducted in Addis Ababa ( 42 ) and Debremarkos ( 19 ) which founds an increase in intention to use family planning as age increases while using family planning for the first time. This could be because, at this age, women are headed toward menopause and the desired number of children is reached, discouraging them from having more children and increasing their intention to use family planning. However, this finding is in contradiction with a study conducted in Aksum ( 43 ) which founds the decrease in intention to use family planning with an increase in age. Similarly, having ever been pregnant increases the likelihood of intending to use family planning. This result was in line with the findings of a study conducted in Adigrat town, Northern Ethiopia ( 19 ) and Aksum ( 43 ). This could be due to; experiencing discomfort and pain during pregnancy and having more children could increase their need for spacing, which, in turn, would increase their intention to use family planning Additionally, having a partner or spouse who is older than 40 years will increase the likelihood of intending to use family planning. The finding is consistent the result of Bangladesh Demographic and Health Survey ( 44 ) and the study done at rural Dembia District ( 45 ). This could be due to the reason that Aged husbands may better share decision-making autonomy with their wives and approve the utilization of modern contraceptive utilization. Husband’s age is also related to better household income which has a positive impact on modern contraceptive utilization. In this study, being single decrease the likelihood of intending to use family planning. This result was consistent with studies conducted in Tanzania( 46 ) and Gondar Town( 47 ). The result highlights the value of male involvement in reproductive health issues, such as fertility and contraception, and couple motivation through education. However, this finding was not compatible with the finding of the study in Mali( 48 ). The two nations' differing socio-demographic compositions and the status of husbands' habits could be the cause. Similarly, having ever been utilized family planning will decrease the likelihood of intending to use family planning among reproductive age women in Ethiopia. Similar finding was documented in studies conducted in Malawi ( 49 ), Gondar ( 50 ) and North Ethiopia (( 51 ). The reasons behind this could include adverse reactions, rejection from spouses, and concerns about the composition of their breast milk changing due to their prior usage of family planning. The priority should thus be on reducing side effects, increasing comfort with use, and ensuring the support of partners for continued use of family planning. In this study need to have more children found to decrease the likelihood of intending to use family planning among reproductive age women in Ethiopia. This finding is supported by the study conducted in the Pakistan ( 52 ), Bangladesh( 53 ) Gondar city( 54 ), and North Shoa Zone ( 55 ). It was clear that women who wanted to have children were not prepared to use family planning. Similarly no unmet need for spacing and limiting was found to decrease the likelihood of intending to use family planning among reproductive age. This finding is in line with the result of cross-sectional study conducted in Ethiopia( 36 ). Unmet needs for family planning arise when a woman of reproductive age, single or married, does not use any kind of birth control despite her desire to stop or postpone having children. Consequently, it follows that a woman who has no unmet needs for spacing and limiting her children's birth may not intend to use family planning. Strength and Limitations of the study The largest problem with utilizing black-box machine learning models, like Random Forest, is that it becomes more difficult to understand the outcomes and the variables influencing the predictions. The researchers employed further investigations to ascertain how factors increased or decreased intention to use family planning to reduce the interpretation limits of machine learning results, which are a result of their black-box nature. In order to evaluate each predictor's relative significance and learn more about how each element affected the model's predictions, the researchers employed a variety of methodologies, including SHAP. Because of this, we were able to comprehend how different factors influenced the model's predictions. Although SHAP explanations are useful for understanding specific forecasts or illustrations, their localized nature limits their ability to provide a global understanding of models. As a result, model behavior or patterns cannot be fully captured by SHAP explanations. Conclusions The results of this investigation showed that reproductive women had low intentions to use family planning. It is therefore highly recommended that the intention of family planning use and initiation of family planning provision should become a standard of service delivery to achieve the 2030 SDGs. This study aimed to identify the major determinants that influence women's intention to utilize family planning. Eight machine learning algorithms were trained to predict the intention to use, and the accuracy and area under the curve of each algorithm were used to assess its performance. The most accurate model, with 77.0% accuracy and 85% areas under the curve, was the random forest model using tenfold cross-validation. Using the SHAP feature importance technique, the top 10 predictors of intention to utilize family planning were found, and their implications were investigated. Lastly, the SHAP analysis showed that having a partner or spouse who is older than 40, being a Muslim, having ever been pregnant, and being older than 30 while using family planning for the first time will increase the likelihood of intending to use family planning. In contrast, the likelihood of intending to use family planning decreases with the following factors: having ever used family planning, being single, having a partner between the ages of 21 and 39, being pregnant, need to have more children, and no unmet need for spacing and limiting. Therefore, family planning providers should emphasize reducing barriers of intention like side effect and discomfort with the family planning method, targeting unmet need for family planning, and the promotion of husband / partner involvement and family planning use. Abbreviations AUC Area under the curve EDHS Ethiopian Demographic and Health Survey KNN K-nearest neighbor LR Logistic regression ML Machine learning RF Random forest ROC Receiver operating characteristic SDG Sustainable Development Goals SMOTE Synthetic Minority Oversampling Technique SSA Sub-Saharan Africa SVM Support Vector Machine WHO World Health Organization XGBoost EXtreme gradient boosting Declarations Ethics approval and consent to participate Permission to use the data has been granted by the PMA Ethiopia’s survey project through legal registration. The data was used which is available on the public domain through the PMA website ( https://www.pmadata.org/data/request-access-datasets ) and can be accessed upon reasonable request after creating an account. Competing interests The authors have declared that no competing interests exist. Contributions JBA, TDN, ADW, DNM, SJW, EBE AND SDK were made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; and agree to be accountable for all aspects of the work. All authors read and approved the final manuscript. Funding The authors received no specific funding for this work. Author Contribution JBA, TDN, ADW, DNM, SJW, EBE AND SDK were made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; and agree to be accountable for all aspects of the work. All authors read and approved the final manuscript. Acknowledgments We are appreciative of the PMA survey project for granting us a license to access the data so that we could carry out the investigation. Without the Python community, this analysis would not have been feasible. Building data analysis pipelines is quick and simple with the many open-source modules and frameworks created by the Python community. This analysis would have been far more challenging and time-consuming without these tools. Data Availability The datasets analyzed in the current study are available in the Performance Monitoring for Action repository, https://www.pmadata.org/data/request-access-datasets . References Bakibinga P, Matanda D, Kisia L, Mutombo N. Factors associated with use of injectables, long-acting and permanent contraceptive methods (iLAPMs) among married women in Zambia: analysis of demographic and health surveys, 1992–2014. Reproductive health. 2019;16(1):1–12. Organization WH. Health in 2015: from MDGs, millennium development goals to SDGs, sustainable development goals. 2015. Boah M, Issah A-N, Demuyakor I, Hyzam D. Long-acting reversible contraceptives utilization and its determinants among married Yemeni women of childbearing age who no longer want children. Medicine. 2022;101(40). Gebeyehu NA, Lake EA, Gelaw KA, Azeze GA. The intention on modern contraceptive use and associated factors among postpartum women in public health institutions of Sodo town, southern Ethiopia 2019: an institutional-based cross-sectional study. BioMed research international. 2020;2020. Lasong J, Zhang Y, Gebremedhin SA, Opoku S, Abaidoo CS, Mkandawire T, et al. Determinants of modern contraceptive use among married women of reproductive age: a cross-sectional study in rural Zambia. BMJ open. 2020;10(3):e030980. Shrivastava SR, Shrivastava PS, Ramasamy J. World Health Organization introduces a new digital tool to address the unmet need for family planning among postpartum women. Sifa Med J. 2016;3(2):60. Cleland J, Bernstein S, Ezeh A, Faundes A, Glasier A, Innis J. Family planning: the unfinished agenda. The lancet. 2006;368(9549):1810–27. Starbird E, Norton M, Marcus R. Investing in family planning: key to achieving the sustainable development goals. Global health: science and practice. 2016;4(2):191–210. Bureau PR. World Population Data Sheet of the Population Reference Bureau, Inc: Demographic Data and Estimates for the Countries and Regions of the World. Population Reference Bureau; 1985. Bongaarts J. The impact of family planning programs on unmet need and demand for contraception. Stud Fam Plann. 2014;45(2):247–62. Tekelab T, Sufa A, Wirtu D. Factors affecting intention to use long acting and permanent contraceptive methods among married women of reproductive age groups in Western Ethiopia: a community based cross sectional study. Fam Med Med Sci Res. 2015;4(158):2. Bulto GA, Zewdie TA, Beyen TK. Demand for long acting and permanent contraceptive methods and associated factors among married women of reproductive age group in Debre Markos Town, North West Ethiopia. BMC Womens Health. 2014;14(1):46. Authority CS. Ethiopia demographic and health survey 20002001. Ahuja M, Frimpong E, Okoro J, Wani R, Armel S. Risk and protective factors for intention of contraception use among women in Ghana. Health Psychol Open. 2020;7(2):2055102920975975. Yzer M. The integrative model of behavioral prediction as a tool for designing health messages. Health communication message design: Theory and practice. 2012;2012:21–40. Lay CH, Burns P. Intentions and behavior in studying for an examination: The role of trait procrastination and its interaction with optimism. J Social Behav Personality. 1991;6(3):605. Peterson HB, Darmstadt GL, Bongaarts J. Meeting the unmet need for family planning: now is the time. The Lancet. 2013;381(9879):1696–9. Gebremariam A, Addissie A. Intention to use long acting and permanent contraceptive methods and factors affecting it among married women in Adigrat town, Tigray, Northern Ethiopia. Reproductive health. 2014;11(1):1–9. Abajobir A. Intention to use long-acting and permanent family planning methods among married 15–49 years women in Debremarkos Town, Northwest Ethiopia. Fam Med Med Sci Res. 2014;3(145):2. Meskele M, Mekonnen W. Factors affecting women’s intention to use long acting and permanent contraceptive methods in Wolaita Zone, Southern Ethiopia: A cross-sectional study. BMC Womens Health. 2014;14:1–9. Abraha TH, Belay HS, Welay GM. Intentions on contraception use and its associated factors among postpartum women in Aksum town, Tigray region, northern Ethiopia: a community-based cross-sectional study. Reproductive health. 2018;15:1–8. Teklab T, Sufa A, Wirtu D. Factors affecting intention to use long-acting and permanent contraceptive methods among married women of reproductive age groups in western Ethiopia: a community based cross-sectional study 2015. Volume 4. Family Medicine & Medical Science Research; 2015. 158. Kebede A, Abaya SG, Merdassa E, Bekuma TT. Factors affecting demand for modern contraceptives among currently married reproductive age women in rural Kebeles of Nunu Kumba district, Oromia, Ethiopia. Contraception and reproductive medicine. 2019;4:1–15. Takele A, Degu G, Yitayal M. Demand for long acting and permanent methods of contraceptives and factors for non-use among married women of Goba Town, Bale Zone, South East Ethiopia. Reproductive health. 2012;9(1):1–11. Alamneh A, Asmamaw A, Woldemariam M, Yenew C, Atikilt G, Andualem M, et al. Trend change in delayed first antenatal care visit among reproductive-aged women in Ethiopia: multivariate decomposition analysis. Reproductive Health. 2022;19(1):1–13. Teshale AB, Tesema GA. Prevalence and associated factors of delayed first antenatal care booking among reproductive age women in Ethiopia; a multilevel analysis of EDHS 2016 data. PLoS ONE. 2020;15(7):e0235538. Ewunetie AA, Munea AM, Meselu BT, Simeneh MM, Meteku BT. DELAY on first antenatal care visit and its associated factors among pregnant women in public health facilities of Debre Markos town, North West Ethiopia. BMC Pregnancy Childbirth. 2018;18(1):1–8. Debelo BT, Danusa KT. Level of Late Initiation of Antenatal Care Visit and Associated Factors Amongst Antenatal Care Attendant Mothers in Gedo General Hospital, West Shoa Zone, Oromia Region, Ethiopia. Front Public Health. 2022;10:866030. Adulo LA, Hassen SS, Chernet A. Timing of the First Antenatal Care Visit and Associated Risk Factors in Rural Parts of Ethiopia. Int J Appl Res Public Health Manage (IJARPHM). 2022;7(1):1–12. Bitew FH, Sparks CS, Nyarko SH. Machine learning algorithms for predicting undernutrition among under-five children in Ethiopia. Public Health Nutr. 2022;25(2):269–80. Goldstein BA, Navar AM, Carter RE. Moving beyond regression techniques in cardiovascular risk prediction: applying machine learning to address analytic challenges. Eur Heart J. 2017;38(23):1805–14. Kebede Kassaw A-A, Melese Yilma T, Sebastian Y, Yeneneh Birhanu A, Sharew Melaku M, Surur Jemal S. Spatial distribution and machine learning prediction of sexually transmitted infections and associated factors among sexually active men and women in Ethiopia, evidence from EDHS 2016. BMC Infect Dis. 2023;23(1):1–16. Kebede SD, Sebastian Y, Yeneneh A, Chanie AF, Melaku MS, Walle AD. Prediction of contraceptive discontinuation among reproductive-age women in Ethiopia using Ethiopian Demographic and Health Survey 2016 Dataset: A Machine Learning Approach. BMC Med Inf Decis Mak. 2023;23(1):1–17. Zimmerman L, Desta S, Yihdego M, Rogers A, Amogne A, Karp C, et al. Protocol for PMA-Ethiopia: A new data source for cross-sectional and longitudinal data of reproductive, maternal, and newborn health. Gates Open Res. 2020;4:126. Shah AD, Bartlett JW, Carpenter J, Nicholas O, Hemingway H. Comparison of random forest and parametric imputation models for imputing missing data using MICE: a CALIBER study. Am J Epidemiol. 2014;179(6):764–74. Kebede SD, Mamo DN, Adem JB, Semagn BE, Walle AD. Machine learning modeling for identifying predictors of unmet need for family planning among married/in-union women in Ethiopia: Evidence from performance monitoring and accountability (PMA) survey 2019 dataset. PLOS Digit Health. 2023;2(10):e0000345. Tiruneh FN, Chuang K-Y, Ntenda PA, Chuang Y-C. Factors associated with contraceptive use and intention to use contraceptives among married women in Ethiopia. Women Health. 2016;56(1):1–22. Weisband Y, Keder L, Keim S, Gallo M. Postpartum intentions on contraceptive use and method choice among breastfeeding women. Contraception. 2016;94(4):405. Eliason S, Baiden F, Quansah-Asare G, Graham-Hayfron Y, Bonsu D, Phillips J, et al. Factors influencing the intention of women in rural Ghana to adopt postpartum family planning. Reproductive health. 2013;10(1):1–8. Gebeyehu NA, Lake EA, Gelaw KA, Azeze GA. The Intention on Modern Contraceptive Use and Associated Factors among Postpartum Women in Public Health Institutions of Sodo Town, Southern Ethiopia 2019: An Institutional-Based Cross-Sectional Study. Biomed Res Int. 2020;2020:9815465. Tegegne BD, Belete MA, Deressa JT. Women’s intention to use long acting and permanent contraceptive methods and associated factors among family planning users in Addis Ababa, Ethiopia: A Cross sectional study. Afr J Reprod Health. 2022;26(4):22–31. Nesro J, Sendo EG, Yesuf NT, Sintayehu Y. Intention to use vasectomy and associated factors among married men in Addis Ababa, Ethiopia. BMC Public Health. 2020;20(1):1228. Syum H, Kahsay G, Huluf T, Beyene B, Gerensea H, Gidey G, et al. Intention to use long-acting and permanent contraceptive methods and associated factors in health institutions of Aksum Town, North Ethiopia. BMC Res Notes. 2019;12(1):739. Hossain M, Khan M, Ababneh F, Shaw JEH. Identifying factors influencing contraceptive use in Bangladesh: evidence from BDHS 2014 data. BMC Public Health. 2018;18(1):1–14. Debebe S, Limenih MA, Biadgo B. Modern contraceptive methods utilization and associated factors among reproductive aged women in rural Dembia District, northwest Ethiopia: Community based cross-sectional study. Int J Reproductive Biomed. 2017;15(6):367. Tengia-Kessy A, Rwabudongo N. Utilization of modern family planning methods among women of reproductive age in a rural setting: the case of Shinyanga rural district, Tanzania. East Afr J Public Heath. 2006;3(2):26–30. Kebede Y. Contraceptive prevalence and factors associated with usage of contraceptives around Gondar Town. Ethiop J Health Dev. 2000;14(3). Kaggwa EB, Diop N, Storey JD. The role of individual and community normative factors: a multilevel analysis of contraceptive use among women in union in Mali. Int Fam Plan Perspect. 2008:79–88. Bwazi C, Maluwa A, Chimwaza A, Pindani M. Utilization of postpartum family planning services between six and twelve months of delivery at Ntchisi District Hospital, Malawi. Health. 2014;2014. Abera Y, Mengesha ZB, Tessema GA. Postpartum contraceptive use in Gondar town, Northwest Ethiopia: a community based cross-sectional study. BMC Womens Health. 2015;15(1):1–8. Abraha TH, Teferra AS, Gelagay AA. Postpartum modern contraceptive use in northern Ethiopia: prevalence and associated factors-methodological issue in this cross-sectional study. Epidemiol health. 2017;39. Stephenson R, Hennink M. Barriers to family planning use amongst the urban poor in Pakistan. 2004. Mostafa Kamal S, Aynul Islam M. Contraceptive use: socioeconomic correlates and method choices in rural Bangladesh. Asia Pac J Public Health. 2010;22(4):436–50. Oumer M, Manaye A, Mengistu Z. Modern contraceptive method utilization and associated factors among women of reproductive age in Gondar City, Northwest Ethiopia. Open Access Journal of Contraception. 2020:53–67. Mohammed A, Woldeyohannes D, Feleke A, Megabiaw B. Determinants of modern contraceptive utilization among married women of reproductive age group in North Shoa Zone, Amhara Region, Ethiopia. Reproductive Health. 2014;11(1):13. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted 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-3848375","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":268753228,"identity":"d966908e-85d7-4a3a-b674-b4686e23e22b","order_by":0,"name":"Jibril Bashir Adem","email":"data:image/png;base64,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","orcid":"","institution":"Arsi University","correspondingAuthor":true,"prefix":"","firstName":"Jibril","middleName":"Bashir","lastName":"Adem","suffix":""},{"id":268753229,"identity":"2c3baa55-36c6-4a76-bffe-05d2aaa64d28","order_by":1,"name":"Tewodros Desalegn Nebi","email":"","orcid":"","institution":"Arsi University","correspondingAuthor":false,"prefix":"","firstName":"Tewodros","middleName":"Desalegn","lastName":"Nebi","suffix":""},{"id":268753230,"identity":"b02fca25-1348-4ff4-bf76-67e9a1802cc6","order_by":2,"name":"Agmasie Damtew Walle","email":"","orcid":"","institution":"Mettu University","correspondingAuthor":false,"prefix":"","firstName":"Agmasie","middleName":"Damtew","lastName":"Walle","suffix":""},{"id":268753231,"identity":"8fb45011-599c-4062-87d1-fefcf2795084","order_by":3,"name":"Daniel Niguse Mamo","email":"","orcid":"","institution":"Arba minch University","correspondingAuthor":false,"prefix":"","firstName":"Daniel","middleName":"Niguse","lastName":"Mamo","suffix":""},{"id":268753232,"identity":"581181cb-fa77-47b1-af12-745e0e061980","order_by":4,"name":"Sudi Jemal Wado","email":"","orcid":"","institution":"Arsi University","correspondingAuthor":false,"prefix":"","firstName":"Sudi","middleName":"Jemal","lastName":"Wado","suffix":""},{"id":268753233,"identity":"a1db9215-eeb5-4eb1-b250-92250dd36d26","order_by":5,"name":"Ermias Belele Enyew","email":"","orcid":"","institution":"Wollo University","correspondingAuthor":false,"prefix":"","firstName":"Ermias","middleName":"Belele","lastName":"Enyew","suffix":""},{"id":268753234,"identity":"38c4f7ab-00c4-410b-adcf-cd8fa87bc9bf","order_by":6,"name":"Shimels Derso Kebede","email":"","orcid":"","institution":"Wollo University","correspondingAuthor":false,"prefix":"","firstName":"Shimels","middleName":"Derso","lastName":"Kebede","suffix":""}],"badges":[],"createdAt":"2024-01-09 13:29:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3848375/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3848375/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":50175613,"identity":"93c6c2ed-9b45-468d-b435-bfea67cc815e","added_by":"auto","created_at":"2024-01-25 16:22:30","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":93232,"visible":true,"origin":"","legend":"\u003cp\u003eOverview flow chart of data preparation and analysis plan applied\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-3848375/v1/94109067736180bc5dd1e09f.png"},{"id":50174752,"identity":"7ae127fa-42b1-4b21-abb4-34e092180f8c","added_by":"auto","created_at":"2024-01-25 16:14:30","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":23761,"visible":true,"origin":"","legend":"\u003cp\u003eRegional distribution of reproductive age women in Ethiopia, 2021\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-3848375/v1/6ac92055a6a3de751dcad789.png"},{"id":50174754,"identity":"ad9a4456-040a-4451-9050-b9633d7d8c16","added_by":"auto","created_at":"2024-01-25 16:14:30","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":45060,"visible":true,"origin":"","legend":"\u003cp\u003eIntention to use family planning among reproductive age women in Ethiopia, 2021\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-3848375/v1/2f36ee7ec973a264693f2a49.png"},{"id":50174753,"identity":"90733f01-daee-4c22-85ea-7407bc047318","added_by":"auto","created_at":"2024-01-25 16:14:30","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":91369,"visible":true,"origin":"","legend":"\u003cp\u003eSHAP global importance plot of optimized Random Forest model\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNote:\u003c/strong\u003e \u003cem\u003efp_ever_used_1 = yes, partner_Age_2 = above 40 years, marital_status_5 = single/unmarried, Religion_4 = Muslim, pregnant_1 = yes, ever_pregnant_1 = yes, more_children_pregnant_2 = No, Unmate_need_for_limiting_1 = no, relationship_3 = son/daughter.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-3848375/v1/499c2c8ae07fd0f1a508edbc.png"},{"id":50174756,"identity":"9c1340fe-599b-4864-9931-017322be9d64","added_by":"auto","created_at":"2024-01-25 16:14:30","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":95139,"visible":true,"origin":"","legend":"\u003cp\u003eBeeswarm plot, ranked by mean absolute SHAP value generated by optimized Random Forest model\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003cstrong\u003eNote\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e: fp_ever_used_1 = yes, partner_Age_2 = above 40 years, marital_status_5 = single/unmarried, Religion_4 = Muslim, pregnant_1 = yes, ever_pregnant_1 = yes, partner_Age_1 = between 21-39 years, more_children_pregnant_2 = yes, Unmate_need_for_limiting_1 = yes, age_at_first_use_30 = above 30 years.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-3848375/v1/35c6283ef4bb61c03e7bb6be.png"},{"id":67201724,"identity":"a8244896-449c-449d-a355-e1f83d85118d","added_by":"auto","created_at":"2024-10-22 10:08:17","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1206218,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3848375/v1/617254d1-8494-4a07-b180-7f2cd0608d4f.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Explainable machine learning algorithm to identify predictors of intention to use family planning among reproductive-age women in Ethiopia: Evidence from the performance monitoring and accountability (PMA) survey 2021 dataset","fulltext":[{"header":"Introduction","content":"\u003cp\u003eIn the next three decades, the population of the world is expected to increase by almost 40% (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). Over half of the additional 3\u0026nbsp;billion people will be in developing regions like sub-Saharan Africa and Asia, where the majority of the population growth is anticipated (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). Out of the 1.9\u0026nbsp;billion women of reproductive age in the globe, 1.1\u0026nbsp;billion need family planning services, and 190\u0026nbsp;million women wish to prevent getting pregnant but do not use any form of contraception(\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eOverall, there is a considerable variation in terms of contraceptive prevalence rate among different methods, ranging from 4.6% in South Sudan to 72.1% in Canada (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). In 2020, approximately 225\u0026nbsp;million people in developing nations wish to delay or cease childbearing, but do not use any form of contraception (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). In the least developed countries, contraceptive usage was significantly lower, at 40%, and was particularly low in Africa at 33% (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). The utilization of family planning is a significant predictor of preventing unwanted pregnancies, reducing maternal and child mortality, and enhancing the wellbeing of women and their families (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Moreover, it is critical to achieving the 2030 Sustainable Development Goals (SDGs) (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eGlobally, the average total fertility rate varies from 1.7 children per woman in the most developed countries to 4.6 in the least developed countries (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e) Within Sub-Saharan Africa (SSA), total fertility rates range from 2.9 in Botswana to 7.2 in Niger (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). The high unmet need for contraception across all age groups and marital statuses accounts for both the rapid population growth and the steady increase in fertility (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWith an estimated 120\u0026nbsp;million people and an average annual growth rate of 2.6%, Ethiopia is the second most populous country in Sub-Saharan Africa (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). By 2050, Ethiopia's population is expected to reach 166\u0026nbsp;million, making it the tenth most populous nation in the world (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). Having a high overall fertility rate, high rates of maternal and infant mortality, and low rates of contraceptive use, Ethiopia is the second most populous country in Sub-Saharan Africa (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTotal fertility rate in Ethiopia is 5.3 children per woman (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). As a result, Ethiopia is now among the nations with the highest global fertility rates. Low family planning utilization is a major contributing factor to the high levels of fertility, particularly in less developed nations like Ethiopia (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). The proportion of married women in Ethiopia who use contraception ranges from 5% in the Somalia region to 53% in Addis Ababa and the Amhara region. The overall modern contraceptive prevalence rate in the country is 40% (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIntention to use and use of family planning has been proven to be an effective method of controlling family size and reducing unintended pregnancies(\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). It is widely believed that intentions are a strong predictor of behavior, and many interventions that aim to change behavior including those targeting contraceptive use rely on evaluating program effectiveness through analyzing behavioral intentions (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). Understanding a woman's intention to use contraceptive methods is crucial in predicting and promoting the use of such methods (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). By doing so, we can improve the health of not just women but also children, families, and even societies (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eNumerous studies conducted on the use and intention to use family planning among reproductive age women have revealed factors such as switching between different forms of contraception (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e), poor support from husband(\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e), maternal age (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e), maternal education (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e), positive attitude to contraceptive use, occupation, knowledge of contraceptives (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e), discussion on family planning with husband(\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e), myths and misconceptions regarding contraception(\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e), time of birth interval(\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e), joint fertility decision(\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e), and desire for live children(\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e) among the variables influencing women of reproductive age's intention to use family planning. Similarly, the intention of family planning use among reproductive age women was significantly associated with, number of live children, and counseling during antenatal care, husbands' approval of family planning use, and having good knowledge of postpartum family planning(\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTraditional regression models were previously used in Ethiopia to demonstrate the effects of socioeconomic, demographic, behavioral, maternal, and service-related characteristics on the intention of reproductive-age women to use family planning; however, the validity of these results declined with an increase in the number of variables and potential correlations (\u003cspan additionalcitationids=\"CR26 CR27 CR28\" citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). The multidisciplinary relationships between variables and numerous factors are typically problematic for these traditional models (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). Hence, in contrast to those traditional models, machine learning (ML) provides an effective way to find relevant characteristics linked to specific health outcomes for conducting public health research (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e). Therefore, the purpose of this study is to evaluate the factors influencing Ethiopian women of reproductive age's intention to use family planning.\u003c/p\u003e \u003cp\u003eThis study seeks to determine and uncover consistent variables as well as new factors that influence the intention to utilize family planning using data from Ethiopian women of reproductive age in the PMA Survey 2021 dataset. The Ethiopian Ministry of Health and other health partners will be able to enhance the intention of reproductive-age women in Ethiopia to utilize family planning by concentrating on the most consistent and influential variables that will be designated as priority areas for intervention.\u003c/p\u003e"},{"header":"Method","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy design\u003c/h2\u003e \u003cp\u003eA machine learning (ML) technique was applied using secondary data from the PMA Ethiopia 2021 cross-sectional household and female Survey. PMA-Ethiopia is a five-year (2019\u0026ndash;2023) project in collaboration with Addis Ababa University, Johns Hopkins University, and the Federal Ministry of Health. It consists of three distinct study activities: yearly cross-sectional surveys of women aged 15\u0026ndash;49; longitudinal surveys of women who are pregnant or have given birth; and yearly service delivery point surveys of health facilities (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eSource and study population\u003c/h2\u003e \u003cp\u003eAll 15\u0026ndash;49 aged women in Ethiopia.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eSample size\u003c/h2\u003e \u003cp\u003eThe sample size in the present study was weighted for taking non-response and variations in the probability of selection into consideration. Further, only women who were of reproductive age at the time of the survey were included in the sample. Therefore, in a weighted sample, only 5993 women of reproductive age were eligible to be included in the analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStudy variables\u003c/h2\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003eDependent variable\u003c/h2\u003e \u003cp\u003eThe dependent variable was the intention to use family planning, which was divided into two groups: \"intended to use\" and \"not intended to use.\"\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003ePredictor variables\u003c/h2\u003e \u003cp\u003eSociodemograpic and economic factors including residence, women's age, region, and religion, size of family, education level, financial status, and media access are predictive of the intention to use family planning. The following reproductive health and family planning service characteristics were also included as predictors of intention to use family planning: partner or husband's feelings toward family planning, knowledge of all available contraceptive methods, age at first sex, having ever been pregnant, having ever used family planning (FP) methods, having ever been delivered in a health facility (HF), and partner being told not to use FP.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eData processing and analysis\u003c/h2\u003e \u003cp\u003eIn order to predict the intention to use family planning, this study employed the general framework found in previous research, which is based on Yufeng Guo's 7 Steps of Machine Learning. The following seven steps in supervised machine learning are outlined in the framework: data preparation, model training, model evaluation, parameter tuning, model selection, data collection, and prediction (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e). Machine learning (ML) algorithms were implemented in Python 3.10.2 using Jupyter Notebook through Scikit-learn, and XGBoost packages.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eData source/collection\u003c/h2\u003e \u003cp\u003eThe dataset for this study is available on the PMA Survey website and can be obtained upon a formal request. A weighted sample of 5993 reproductive-age women is included in the study.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eData preparation/pre-processing\u003c/h2\u003e \u003cp\u003eThis study employed a variety of data preparation techniques, including feature engineering, data splitting, and data cleaning. Missing values were imputed using the R package 'CALIBERrfimpute' using Multivariate Imputation by Chained Equations (MICE) and random forest and the training data was balanced using the Synthetic Minority Oversampling Technique (SMOTE)(\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). SMOTE creates synthetic observations by interpolating between minority classes samples in the feature space. The Kolmogorov-Smirnov test indicates that the synthetic and existing observations differ significantly from each other (KS test statistic of 0.8214, p-value\u0026thinsp;=\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Using Pandas get-dummies, One-Hot-Encoding techniques were applied to encode categorical data into dummy variables, and the relationship between the predictors and the outcome variable was evaluated using the SHAP feature importance method. A tenfold cross-validation technique was used to train the models(\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eModel selection and training\u003c/h2\u003e \u003cp\u003eWe used eight different machine learning algorithms namely AdaBoost, logistic regression (LR), random forest (RF), K-nearest neighbor (KNN), artificial neural network, support vector machine (SVM), Na\u0026iuml;ve Bayes, and extreme gradient boosting (XGBoost) to predict intention to use family planning among reproductive age women and then determined which model was best suited for the problem(\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). Tenfold cross-validation was used to compare the performance of the chosen classifiers, and the best predictive model was chosen after comparison, adjusted with its hyper-parameters, and trained using balanced training data.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eModel evaluation and Hyper-parameter tuning\u003c/h2\u003e \u003cp\u003eFollowing training, the model was evaluated using the area under the receiver operating characteristic curve (AUC) score and classification accuracy. Moreover, ML model performance was shown using the receiver operating characteristic (ROC) curve. The Optuna framework was used to adjust the best model's hyper-parameters. It uses the Bayesian framework to better understand the probability of the optimal values and prevents needless computation for the combination of non-performing parameters in the search(\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003ePrediction\u003c/h2\u003e \u003cp\u003ePrediction involves estimating the result variable using pre-selected predictors and applying the fully trained model to its intended output. In this study, the intention to utilize family planning was ascertained using the top predictors identified during the analysis. Consequently, the most accurate classifier will determine, to a predetermined degree of accuracy, whether or not a woman intends to use family planning.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eModel interpretation/explanation using Shapley Additive exPlanations (SHAP)\u003c/h2\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAn interpretation and explanation of any machine learning model's prediction can be found globally or locally using Shapley Additive exPlanations (SHAP) analysis. The fundamental concept behind the SHAP analysis is figuring out each predictor's marginal contribution to the outcome variable prediction result. In machine learning, features are chosen using SHAP values. The aggregate Shapley value of each characteristic for each sample should be shown in order to assess the impact of each predictor on the prediction of intention to use family planning. Moreover, a waterfall plot was used to illustrate how each component contributed to the prediction of a positive class. Finally, the overall data preparation and analysis process is presented in (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). A thorough explanation of the SHAP analysis can be found in the previous article (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eSocio-demographic and economic characteristics\u003c/h2\u003e \u003cp\u003eThe study included a weighted sample of 5993 women who were of reproductive age. The mean age of the sample was 23.89 years (\u0026plusmn;\u0026thinsp;19.2 standard deviation), with the majority of the women, 5,898 (71.87%), ranging between the ages of 15 and 19. 5120 (63.42%) of the participants were married or living with their spouses, while 4,957 (60.41%) were rural residents. Muslims made up 31.91% of the attendees, while Orthodox Christians made up the majority with 3,068 (38.00%). The majority of women, 2,595 (32.14%), have completed secondary school or above (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003esocio-demographic status of reproductive age women in Ethiopia, 2021\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCategories\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWeighted Freq.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePercent (%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eResidence\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUrban\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,249\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e39.59\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRural\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4,957\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e60.41\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eMarital status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMarried\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5,120\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e63.42\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003esingle\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2,220\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e27.50\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDivorced/widowed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e733\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e9.08\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eWomen\u0026rsquo;s age\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15 to 19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5,898\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e71.87\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20 to 29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1,317\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e16.05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30 to 39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e712\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e8.68\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e40 to 49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e279\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.40\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eEducation level\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2,421\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e29.99\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePrimary\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,057\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e37.87\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003esecondary and above\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2,595\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e32.14\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eWealth status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePoor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2,742\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e33.42\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMiddle\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1,386\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e16.89\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRich\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4,077\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e49.69\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eMedia access\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5,458\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e67.78\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2,594\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e32.22\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eReligion\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOrthodox\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,068\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e38.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMuslim\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2,576\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e31.91\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eProtestant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2,363\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e29.27\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOthers*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.82\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eNote: \u003cb\u003eothers*\u003c/b\u003e indicates: catholic, Wakefata, atheist and traditional\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eRegarding the regional distribution of respondents, The Amhara region 1,584(19.30%), the Southern Nations, Nationalities, and Peoples' region 1,322 (16.11%), and the Oromia region accounted for the majority of respondents (1,858 (22.64%). No individuals from the Tigray regional state were in the study due to the conflict in the area throughout the study period (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eReproductive health and Family planning service characteristics\u003c/h2\u003e \u003cp\u003eOf the total respondents, 5,381 (60.41%) had previously been pregnant, and about 11.82% of the women (49.13%) had unmet family planning needs. About (48.3%) of the respondents had never utilized any family planning techniques. approximately 4,540 (74.69%) of the women started experiencing sex between the ages of 15 and 19, and 5.85% of the individuals said that their partner forced them into getting pregnant (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eReproductive health and family planning service characteristics of study participants\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCategories\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWeighted Freq.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePercent (%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eEver been pregnant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2,691\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e33.34\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5,381\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e60.41\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eEver used family planning\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,876\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e48.31\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4,148\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e51.69\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eVisited a facility for care for self or children in past 12 months\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4,184\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e51.96\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,868\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e48.04\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eCurrent used of any contraceptive methods\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5,933\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e74.27\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2,055\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e25.73\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eunmet need for family planning\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7,116\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e88.18\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e954\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e11.82\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eYour partner tried to force you to become pregnant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4,815\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e94.15\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e299\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5.85\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eWas visited by HEW who talked about FP in past 12 months\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7,370\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e91.59\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e677\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e8.41\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eAge at first sexual intercourse\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15 to 19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4,539\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e74.69\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20 to 29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1,478\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e24.32\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30 to 39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.91\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e40 to 49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eIntention to use family planning among reproductive age women in Ethiopia, based on PMA survey, 2021\u003c/b\u003e \u003c/p\u003e \u003cp\u003eAmong the study participants, 2,891(48.24%) reproductive age women had no intention of using any form of family planning (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eMachine learning approach to identify predictors of intention to use family planning among reproductive age women in Ethiopia, 2021 PMA survey\u003c/b\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eBalancing data and Model performance comparison\u003c/h2\u003e \u003cp\u003eIn order to balance the outcome variable's unequal distribution, 168 additional synthetic observations from the minority category those who did not intend to utilize family planning were produced by the SMOTE oversampling approach. In order to create symmetric distributions for both categories and enable the development of reliable prediction models, the total intention to use status distribution was adjusted from 2313 not intended to use and 2481 intended to use, resulting in 2481 in each class.\u003c/p\u003e \u003cp\u003eTo assess the performance of the models to predict intention family planning among reproductive age women, a stratified 10-fold cross-validation was employed, with particular consideration paid to mean accuracy and mean area under the curve score. Stratified 10-fold cross-validation was used to assess classifiers on the unbalanced training data, and the support vector machine emerged to the top with an accuracy of 76.7% and 84.2% areas under the ROC curve. The unbalanced class character of the outcome variable, which could skew the model in favor of the majority class, makes this finding potentially unreliable. After balancing the training data using the SMOTE oversampling technique, an ML model comparison was conducted in order to prevent this biased model creation. Accordingly, with a 76.8% accuracy rate and an 84.6% area under the ROC curve, Random Forest emerged as the most accurate predictive model to predict intention to use family planning among reproductive age women in Ethiopia (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eModel comparison through tenfold cross-validation on training data\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModels\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePerformance\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUnbalanced data\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eBalanced data\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eLogistic regression\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e76.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e75.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e83.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e833\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSVM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e76.7*\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e76.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e84.2*\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e84.3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eXGBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e75.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e76.3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e82.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e83.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eKNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e72.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e72.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e78.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e78.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eRandom forest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e76.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e76.8*\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e83.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e84.6*\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eAdaBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e75.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e75.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e82.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e83.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eNa\u0026iuml;ve Bayes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e72.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e71.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e77.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e77.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eArtificial neural network\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e72.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e73.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e79.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e80.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eHyper-parameter tuning of Random Forest\u003c/h2\u003e \u003cp\u003eScikit-learn do not always provide the best solution for a given problem, even if it offers a set of reasonable default hyper-parameters for all models, including Random Forest. The number of decision trees in the forest (n_estimators), the number of features that each tree considers when splitting a node (max_features), the minimum number of samples needed to split an internal node (min_samples_split), the minimum number of samples needed to be at a leaf node (min_samples_leaf), and the number of samples to draw from independent variables to train each tree (max_samples) were among the hyper-parameters that were optimized with one hundred seventy five trials on a specified search space using stratified 10-fold cross-validation in order to maximize the performance of random forests. The Scikit-learn default hyper-parameters and our adjusted hyper-parameters are displayed in (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDefault and optimal tuned hyper-parameters of Random Forest model\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHyper-parameters\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDefault\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOptimal value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNumber of tree\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e175\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNumber of features considered for the best split\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSquare root of the number of features\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.302\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eminimum number of samples required to split an internal node\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eminimum number of samples required to be at a leaf node\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003enumber of samples to draw from independent variables to train each base estimator\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNone\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eFinally, random forest model was created with these tuned hyper-parameters on balanced training data through 10-fold cross-validation and yielded 77.0% accuracy and 85% areas under the curve.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eFeature selection\u003c/h2\u003e \u003cp\u003eFollowing the selection of the best model, previously unseen test data was used to predict the intention to utilize family planning. Following random forest training on unbalanced training data, balanced data using default model parameters, and a comparison with an optimized model trained on balanced data, the prediction was made. Following balanced and unbalanced data training for the random forest model, the prediction on unseen test data yielded an area under the ROC curve score of 0.83 for both. An improved AUC of 0.85 was anticipated by hyper-parameter-tuned random forest, nevertheless. In this study, the top predictors of the intention to use family planning were determined using model-agnostic SHAP global feature importance. In order to quantify a feature's contribution to the anticipated intention to use family planning among reproductive age women in Ethiopia, this technique looks at the mean absolute SHAP value for each predictor over all of the data. Higher mean absolute SHAP values indicate greater influence of the predictors, which are arranged in descending order of their impact on the outcome variable prediction.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe findings showed that the most significant predictors of women's intention to use family planning were ever been utilized family planning (fp_ever_used_1), a partner or husband older than 40 (partner age_2), being single (marital_status_5), and being Muslim (religion_4). Other significant predictors of intention to use family planning included being pregnant (pregnant_1), having ever been pregnant (ever_pregnant_1), need to have more children (more_children_pregnant_2), having a son or daughter relationship to the head of the household (relationship_3), and unmet need for spacing and limiting. Each class's horizontal rectangle is half-filled with the colors red and blue, as it\u0026rsquo;s shown in the Fig.\u0026nbsp;6. As a result, each feature has an equal influence on whether a case is classified as intended to use family planning (label\u0026thinsp;=\u0026thinsp;1) or not intended to use family planning (label\u0026thinsp;=\u0026thinsp;0) (Fig.\u0026nbsp;4).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eModel interpretation/explanation\u003c/h2\u003e \u003cp\u003eA comprehensive picture of how the factors affect the model's predictions across the board was shown by using Beeswarms plots. Plotting the Shapley value of each individual predictor for each sample, Fig.\u0026nbsp;7 shows the distribution of the impacts of each predictor on the model's output (i.e., intention to use family planning prediction).\u003c/p\u003e \u003cp\u003eThe points on this beeswarm plot represent Shapley values of the features related to intention to use family planning status, providing insight into the importance and association of each of the top ten features on the outcome variable. The red and blue hues in the figure represent the higher and lower values of each predictor\u0026rsquo;s variable. Points that are right to the vertical line (0 SHAP value) increase the likelihood of intention to use family planning while the left side decreases likelihood of intention to use family.\u003c/p\u003e \u003cp\u003eSince all of the variables are categorical and have two categories, the blue line denotes category code 0 (low value) and the red line denotes category code 1, which is a high value. Therefore, having a partner or spouse who is older than 40 (partner_Age_2), being a Muslim (Religion_4), having ever been pregnant (ever_pregnant_1), and being older than 30 while using family planning for the first time will increase the likelihood of intending to use family planning.\u003c/p\u003e \u003cp\u003eContrarily, the likelihood of intending to use family planning decreases with the following factors: having ever used family planning (fp_ever_used_1), being single (marital_status_5), having a partner between the ages of 21 and 39 (partner_age_1), being pregnant (pregnant_1), need to have more children (more_children_pregnant_2), and no unmet need for spacing and limiting (Unmate_need_for_spacing_limiting_1) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eTenfold cross-validation was used to train eight machine learning classifiers on balanced and imbalanced training data. Random Forest was found to be the most accurate predictive model, and it was optimized for its optimal hyper-parameters before more analysis was carried out. Based on the result of this study, 51% of reproductive age women in Ethiopia had intention of using any form of family planning. This finding is higher than the result of a study conducted on Ethiopian Demographic and Health Survey, 2011 (44.1%) (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e) but lower than the result of studies conducted in 90%(\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e), Gahanna (\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e) and Sodo(\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e) the prevalence was found to be 70%, Addis Ababa 60% (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e). The reason behind the lower magnitude of intention to use contraception was the difference in availability, accessibility, infrastructure, and socioeconomic status between the present study, and a study done in America.\u003c/p\u003e \u003cp\u003eThe SHAP analysis revealed that age at first family planning use, a partner or husband older than 40, being single, being Muslim, being pregnant, having previously been pregnant, need to have more children, having a son or daughter relationship to the head of the household and need for spacing and limiting were significant predictors.\u003c/p\u003e \u003cp\u003eAccording to the SHAP model explanation being older than 30 while using family planning for the first time will increase the likelihood of intending to use family planning. This finding is in line with the finding of study conducted in Addis Ababa (\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e) and Debremarkos (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e) which founds an increase in intention to use family planning as age increases while using family planning for the first time. This could be because, at this age, women are headed toward menopause and the desired number of children is reached, discouraging them from having more children and increasing their intention to use family planning. However, this finding is in contradiction with a study conducted in Aksum (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e) which founds the decrease in intention to use family planning with an increase in age.\u003c/p\u003e \u003cp\u003eSimilarly, having ever been pregnant increases the likelihood of intending to use family planning. This result was in line with the findings of a study conducted in Adigrat town, Northern Ethiopia (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e) and Aksum (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e). This could be due to; experiencing discomfort and pain during pregnancy and having more children could increase their need for spacing, which, in turn, would increase their intention to use family planning\u003c/p\u003e \u003cp\u003eAdditionally, having a partner or spouse who is older than 40 years will increase the likelihood of intending to use family planning. The finding is consistent the result of Bangladesh Demographic and Health Survey (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e) and the study done at rural Dembia District (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e). This could be due to the reason that Aged husbands may better share decision-making autonomy with their wives and approve the utilization of modern contraceptive utilization. Husband\u0026rsquo;s age is also related to better household income which has a positive impact on modern contraceptive utilization.\u003c/p\u003e \u003cp\u003eIn this study, being single decrease the likelihood of intending to use family planning. This result was consistent with studies conducted in Tanzania(\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e) and Gondar Town(\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e). The result highlights the value of male involvement in reproductive health issues, such as fertility and contraception, and couple motivation through education. However, this finding was not compatible with the finding of the study in Mali(\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e). The two nations' differing socio-demographic compositions and the status of husbands' habits could be the cause.\u003c/p\u003e \u003cp\u003eSimilarly, having ever been utilized family planning will decrease the likelihood of intending to use family planning among reproductive age women in Ethiopia. Similar finding was documented in studies conducted in Malawi (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e), Gondar (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e) and North Ethiopia ((\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e). The reasons behind this could include adverse reactions, rejection from spouses, and concerns about the composition of their breast milk changing due to their prior usage of family planning. The priority should thus be on reducing side effects, increasing comfort with use, and ensuring the support of partners for continued use of family planning.\u003c/p\u003e \u003cp\u003eIn this study need to have more children found to decrease the likelihood of intending to use family planning among reproductive age women in Ethiopia. This finding is supported by the study conducted in the Pakistan (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e), Bangladesh(\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e) Gondar city(\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e), and North Shoa Zone (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e). It was clear that women who wanted to have children were not prepared to use family planning.\u003c/p\u003e \u003cp\u003eSimilarly no unmet need for spacing and limiting was found to decrease the likelihood of intending to use family planning among reproductive age. This finding is in line with the result of cross-sectional study conducted in Ethiopia(\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). Unmet needs for family planning arise when a woman of reproductive age, single or married, does not use any kind of birth control despite her desire to stop or postpone having children. Consequently, it follows that a woman who has no unmet needs for spacing and limiting her children's birth may not intend to use family planning.\u003c/p\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eStrength and Limitations of the study\u003c/h2\u003e \u003cp\u003eThe largest problem with utilizing black-box machine learning models, like Random Forest, is that it becomes more difficult to understand the outcomes and the variables influencing the predictions. The researchers employed further investigations to ascertain how factors increased or decreased intention to use family planning to reduce the interpretation limits of machine learning results, which are a result of their black-box nature. In order to evaluate each predictor's relative significance and learn more about how each element affected the model's predictions, the researchers employed a variety of methodologies, including SHAP. Because of this, we were able to comprehend how different factors influenced the model's predictions. Although SHAP explanations are useful for understanding specific forecasts or illustrations, their localized nature limits their ability to provide a global understanding of models. As a result, model behavior or patterns cannot be fully captured by SHAP explanations.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThe results of this investigation showed that reproductive women had low intentions to use family planning. It is therefore highly recommended that the intention of family planning use and initiation of family planning provision should become a standard of service delivery to achieve the 2030 SDGs.\u003c/p\u003e \u003cp\u003eThis study aimed to identify the major determinants that influence women's intention to utilize family planning. Eight machine learning algorithms were trained to predict the intention to use, and the accuracy and area under the curve of each algorithm were used to assess its performance. The most accurate model, with 77.0% accuracy and 85% areas under the curve, was the random forest model using tenfold cross-validation.\u003c/p\u003e \u003cp\u003eUsing the SHAP feature importance technique, the top 10 predictors of intention to utilize family planning were found, and their implications were investigated. Lastly, the SHAP analysis showed that having a partner or spouse who is older than 40, being a Muslim, having ever been pregnant, and being older than 30 while using family planning for the first time will increase the likelihood of intending to use family planning. In contrast, the likelihood of intending to use family planning decreases with the following factors: having ever used family planning, being single, having a partner between the ages of 21 and 39, being pregnant, need to have more children, and no unmet need for spacing and limiting. Therefore, family planning providers should emphasize reducing barriers of intention like side effect and discomfort with the family planning method, targeting unmet need for family planning, and the promotion of husband / partner involvement and family planning use.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eAUC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eArea under the curve\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eEDHS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eEthiopian Demographic and Health Survey\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eKNN\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eK-nearest neighbor\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eLogistic regression\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eML\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eMachine learning\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRF\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eRandom forest\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eROC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eReceiver operating characteristic\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSDG\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eSustainable Development Goals\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSMOTE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eSynthetic Minority Oversampling Technique\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSSA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eSub-Saharan Africa\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSVM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eSupport Vector Machine\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eWHO\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eWorld Health Organization\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eXGBoost\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eEXtreme gradient boosting\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003ch2\u003eEthics approval and consent to participate\u003c/h2\u003e \u003cp\u003ePermission to use the data has been granted by the PMA Ethiopia\u0026rsquo;s survey project through legal registration. The data was used which is available on the public domain through the PMA website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.pmadata.org/data/request-access-datasets\u003c/span\u003e\u003cspan address=\"https://www.pmadata.org/data/request-access-datasets\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and can be accessed upon reasonable request after creating an account.\u003c/p\u003e \u003ch2\u003eCompeting interests\u003c/h2\u003e \u003cp\u003eThe authors have declared that no competing interests exist.\u003c/p\u003e\u003ch2\u003eContributions\u003c/h2\u003e \u003cp\u003e JBA, TDN, ADW, DNM, SJW, EBE AND SDK were made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; and agree to be accountable for all aspects of the work. All authors read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThe authors received no specific funding for this work.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eJBA, TDN, ADW, DNM, SJW, EBE AND SDK were made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; and agree to be accountable for all aspects of the work. All authors read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgments\u003c/h2\u003e \u003cp\u003eWe are appreciative of the PMA survey project for granting us a license to access the data so that we could carry out the investigation. Without the Python community, this analysis would not have been feasible. Building data analysis pipelines is quick and simple with the many open-source modules and frameworks created by the Python community. This analysis would have been far more challenging and time-consuming without these tools.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e \u003cp\u003eThe datasets analyzed in the current study are available in the Performance Monitoring for Action repository, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.pmadata.org/data/request-access-datasets\u003c/span\u003e\u003cspan address=\"https://www.pmadata.org/data/request-access-datasets\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBakibinga P, Matanda D, Kisia L, Mutombo N. Factors associated with use of injectables, long-acting and permanent contraceptive methods (iLAPMs) among married women in Zambia: analysis of demographic and health surveys, 1992\u0026ndash;2014. Reproductive health. 2019;16(1):1\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOrganization WH. Health in 2015: from MDGs, millennium development goals to SDGs, sustainable development goals. 2015.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoah M, Issah A-N, Demuyakor I, Hyzam D. Long-acting reversible contraceptives utilization and its determinants among married Yemeni women of childbearing age who no longer want children. Medicine. 2022;101(40).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGebeyehu NA, Lake EA, Gelaw KA, Azeze GA. The intention on modern contraceptive use and associated factors among postpartum women in public health institutions of Sodo town, southern Ethiopia 2019: an institutional-based cross-sectional study. BioMed research international. 2020;2020.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLasong J, Zhang Y, Gebremedhin SA, Opoku S, Abaidoo CS, Mkandawire T, et al. Determinants of modern contraceptive use among married women of reproductive age: a cross-sectional study in rural Zambia. BMJ open. 2020;10(3):e030980.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShrivastava SR, Shrivastava PS, Ramasamy J. World Health Organization introduces a new digital tool to address the unmet need for family planning among postpartum women. Sifa Med J. 2016;3(2):60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCleland J, Bernstein S, Ezeh A, Faundes A, Glasier A, Innis J. Family planning: the unfinished agenda. The lancet. 2006;368(9549):1810\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStarbird E, Norton M, Marcus R. Investing in family planning: key to achieving the sustainable development goals. Global health: science and practice. 2016;4(2):191\u0026ndash;210.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBureau PR. World Population Data Sheet of the Population Reference Bureau, Inc: Demographic Data and Estimates for the Countries and Regions of the World. Population Reference Bureau; 1985.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBongaarts J. The impact of family planning programs on unmet need and demand for contraception. Stud Fam Plann. 2014;45(2):247\u0026ndash;62.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTekelab T, Sufa A, Wirtu D. Factors affecting intention to use long acting and permanent contraceptive methods among married women of reproductive age groups in Western Ethiopia: a community based cross sectional study. Fam Med Med Sci Res. 2015;4(158):2.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBulto GA, Zewdie TA, Beyen TK. Demand for long acting and permanent contraceptive methods and associated factors among married women of reproductive age group in Debre Markos Town, North West Ethiopia. BMC Womens Health. 2014;14(1):46.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAuthority CS. Ethiopia demographic and health survey 20002001.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAhuja M, Frimpong E, Okoro J, Wani R, Armel S. Risk and protective factors for intention of contraception use among women in Ghana. Health Psychol Open. 2020;7(2):2055102920975975.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYzer M. The integrative model of behavioral prediction as a tool for designing health messages. Health communication message design: Theory and practice. 2012;2012:21\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLay CH, Burns P. Intentions and behavior in studying for an examination: The role of trait procrastination and its interaction with optimism. J Social Behav Personality. 1991;6(3):605.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeterson HB, Darmstadt GL, Bongaarts J. Meeting the unmet need for family planning: now is the time. The Lancet. 2013;381(9879):1696\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGebremariam A, Addissie A. Intention to use long acting and permanent contraceptive methods and factors affecting it among married women in Adigrat town, Tigray, Northern Ethiopia. Reproductive health. 2014;11(1):1\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbajobir A. Intention to use long-acting and permanent family planning methods among married 15\u0026ndash;49 years women in Debremarkos Town, Northwest Ethiopia. Fam Med Med Sci Res. 2014;3(145):2.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeskele M, Mekonnen W. Factors affecting women\u0026rsquo;s intention to use long acting and permanent contraceptive methods in Wolaita Zone, Southern Ethiopia: A cross-sectional study. BMC Womens Health. 2014;14:1\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbraha TH, Belay HS, Welay GM. Intentions on contraception use and its associated factors among postpartum women in Aksum town, Tigray region, northern Ethiopia: a community-based cross-sectional study. Reproductive health. 2018;15:1\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeklab T, Sufa A, Wirtu D. Factors affecting intention to use long-acting and permanent contraceptive methods among married women of reproductive age groups in western Ethiopia: a community based cross-sectional study 2015. Volume 4. Family Medicine \u0026amp; Medical Science Research; 2015. 158.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKebede A, Abaya SG, Merdassa E, Bekuma TT. Factors affecting demand for modern contraceptives among currently married reproductive age women in rural Kebeles of Nunu Kumba district, Oromia, Ethiopia. Contraception and reproductive medicine. 2019;4:1\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTakele A, Degu G, Yitayal M. Demand for long acting and permanent methods of contraceptives and factors for non-use among married women of Goba Town, Bale Zone, South East Ethiopia. Reproductive health. 2012;9(1):1\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlamneh A, Asmamaw A, Woldemariam M, Yenew C, Atikilt G, Andualem M, et al. Trend change in delayed first antenatal care visit among reproductive-aged women in Ethiopia: multivariate decomposition analysis. Reproductive Health. 2022;19(1):1\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeshale AB, Tesema GA. Prevalence and associated factors of delayed first antenatal care booking among reproductive age women in Ethiopia; a multilevel analysis of EDHS 2016 data. PLoS ONE. 2020;15(7):e0235538.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEwunetie AA, Munea AM, Meselu BT, Simeneh MM, Meteku BT. DELAY on first antenatal care visit and its associated factors among pregnant women in public health facilities of Debre Markos town, North West Ethiopia. BMC Pregnancy Childbirth. 2018;18(1):1\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDebelo BT, Danusa KT. Level of Late Initiation of Antenatal Care Visit and Associated Factors Amongst Antenatal Care Attendant Mothers in Gedo General Hospital, West Shoa Zone, Oromia Region, Ethiopia. Front Public Health. 2022;10:866030.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAdulo LA, Hassen SS, Chernet A. Timing of the First Antenatal Care Visit and Associated Risk Factors in Rural Parts of Ethiopia. Int J Appl Res Public Health Manage (IJARPHM). 2022;7(1):1\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBitew FH, Sparks CS, Nyarko SH. Machine learning algorithms for predicting undernutrition among under-five children in Ethiopia. Public Health Nutr. 2022;25(2):269\u0026ndash;80.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoldstein BA, Navar AM, Carter RE. Moving beyond regression techniques in cardiovascular risk prediction: applying machine learning to address analytic challenges. Eur Heart J. 2017;38(23):1805\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKebede Kassaw A-A, Melese Yilma T, Sebastian Y, Yeneneh Birhanu A, Sharew Melaku M, Surur Jemal S. Spatial distribution and machine learning prediction of sexually transmitted infections and associated factors among sexually active men and women in Ethiopia, evidence from EDHS 2016. BMC Infect Dis. 2023;23(1):1\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKebede SD, Sebastian Y, Yeneneh A, Chanie AF, Melaku MS, Walle AD. Prediction of contraceptive discontinuation among reproductive-age women in Ethiopia using Ethiopian Demographic and Health Survey 2016 Dataset: A Machine Learning Approach. BMC Med Inf Decis Mak. 2023;23(1):1\u0026ndash;17.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZimmerman L, Desta S, Yihdego M, Rogers A, Amogne A, Karp C, et al. Protocol for PMA-Ethiopia: A new data source for cross-sectional and longitudinal data of reproductive, maternal, and newborn health. Gates Open Res. 2020;4:126.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShah AD, Bartlett JW, Carpenter J, Nicholas O, Hemingway H. Comparison of random forest and parametric imputation models for imputing missing data using MICE: a CALIBER study. Am J Epidemiol. 2014;179(6):764\u0026ndash;74.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKebede SD, Mamo DN, Adem JB, Semagn BE, Walle AD. Machine learning modeling for identifying predictors of unmet need for family planning among married/in-union women in Ethiopia: Evidence from performance monitoring and accountability (PMA) survey 2019 dataset. PLOS Digit Health. 2023;2(10):e0000345.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTiruneh FN, Chuang K-Y, Ntenda PA, Chuang Y-C. Factors associated with contraceptive use and intention to use contraceptives among married women in Ethiopia. Women Health. 2016;56(1):1\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeisband Y, Keder L, Keim S, Gallo M. Postpartum intentions on contraceptive use and method choice among breastfeeding women. Contraception. 2016;94(4):405.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEliason S, Baiden F, Quansah-Asare G, Graham-Hayfron Y, Bonsu D, Phillips J, et al. Factors influencing the intention of women in rural Ghana to adopt postpartum family planning. Reproductive health. 2013;10(1):1\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGebeyehu NA, Lake EA, Gelaw KA, Azeze GA. The Intention on Modern Contraceptive Use and Associated Factors among Postpartum Women in Public Health Institutions of Sodo Town, Southern Ethiopia 2019: An Institutional-Based Cross-Sectional Study. Biomed Res Int. 2020;2020:9815465.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTegegne BD, Belete MA, Deressa JT. Women\u0026rsquo;s intention to use long acting and permanent contraceptive methods and associated factors among family planning users in Addis Ababa, Ethiopia: A Cross sectional study. Afr J Reprod Health. 2022;26(4):22\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNesro J, Sendo EG, Yesuf NT, Sintayehu Y. Intention to use vasectomy and associated factors among married men in Addis Ababa, Ethiopia. BMC Public Health. 2020;20(1):1228.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSyum H, Kahsay G, Huluf T, Beyene B, Gerensea H, Gidey G, et al. Intention to use long-acting and permanent contraceptive methods and associated factors in health institutions of Aksum Town, North Ethiopia. BMC Res Notes. 2019;12(1):739.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHossain M, Khan M, Ababneh F, Shaw JEH. Identifying factors influencing contraceptive use in Bangladesh: evidence from BDHS 2014 data. BMC Public Health. 2018;18(1):1\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDebebe S, Limenih MA, Biadgo B. Modern contraceptive methods utilization and associated factors among reproductive aged women in rural Dembia District, northwest Ethiopia: Community based cross-sectional study. Int J Reproductive Biomed. 2017;15(6):367.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTengia-Kessy A, Rwabudongo N. Utilization of modern family planning methods among women of reproductive age in a rural setting: the case of Shinyanga rural district, Tanzania. East Afr J Public Heath. 2006;3(2):26\u0026ndash;30.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKebede Y. Contraceptive prevalence and factors associated with usage of contraceptives around Gondar Town. Ethiop J Health Dev. 2000;14(3).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKaggwa EB, Diop N, Storey JD. The role of individual and community normative factors: a multilevel analysis of contraceptive use among women in union in Mali. Int Fam Plan Perspect. 2008:79\u0026ndash;88.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBwazi C, Maluwa A, Chimwaza A, Pindani M. Utilization of postpartum family planning services between six and twelve months of delivery at Ntchisi District Hospital, Malawi. Health. 2014;2014.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbera Y, Mengesha ZB, Tessema GA. Postpartum contraceptive use in Gondar town, Northwest Ethiopia: a community based cross-sectional study. BMC Womens Health. 2015;15(1):1\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbraha TH, Teferra AS, Gelagay AA. Postpartum modern contraceptive use in northern Ethiopia: prevalence and associated factors-methodological issue in this cross-sectional study. Epidemiol health. 2017;39.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStephenson R, Hennink M. Barriers to family planning use amongst the urban poor in Pakistan. 2004.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMostafa Kamal S, Aynul Islam M. Contraceptive use: socioeconomic correlates and method choices in rural Bangladesh. Asia Pac J Public Health. 2010;22(4):436\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOumer M, Manaye A, Mengistu Z. Modern contraceptive method utilization and associated factors among women of reproductive age in Gondar City, Northwest Ethiopia. Open Access Journal of Contraception. 2020:53\u0026ndash;67.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMohammed A, Woldeyohannes D, Feleke A, Megabiaw B. Determinants of modern contraceptive utilization among married women of reproductive age group in North Shoa Zone, Amhara Region, Ethiopia. Reproductive Health. 2014;11(1):13.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Machine learning algorithm, Intention to use family planning, Predictors, Ethiopia","lastPublishedDoi":"10.21203/rs.3.rs-3848375/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3848375/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eIntroduction:\u003c/h2\u003e \u003cp\u003eApproximately 225\u0026nbsp;million people in developing nations wish to delay or cease childbearing, but do not use any form of contraception. In the least developed countries, contraceptive usage was significantly lower, at 40%, and was particularly low in Africa at 33%. It is widely believed that intentions are a strong predictor of behavior, and many interventions that aim to change behavior including that targeting family planning use rely on evaluating program effectiveness through analyzing behavioral intentions. Understanding a woman's intention to use contraceptive methods is crucial in predicting and promoting the use of such methods. Therefore, this study aims to assess the determinants of intention to use family planning among reproductive age women in Ethiopia using explainable machine learning algorithm\u003c/p\u003e\u003ch2\u003eMethod\u003c/h2\u003e \u003cp\u003eSecondary data from the cross-sectional household and female survey conducted by PMA Ethiopia in 2021 were used in the study. Using Python 3.10 version software, eight machine learning classifiers were used to predict and identify significant determinants of intention to use family planning on a weighted sample of 5993 women. Performance metrics were used to evaluate the classifiers. To smooth the data for additional analysis, data preparation techniques such as feature engineering, data splitting, handling missing values, addressing imbalanced categories, and outlier removal were used. Lastly, the greatest predictors of intention to utilize family planning were found using Shapley Additive exPlanations (SHAP) analysis, which further clarified the predictors' impact on the model's results.\u003c/p\u003e\u003ch2\u003eResult\u003c/h2\u003e \u003cp\u003eUsing tenfold cross-validation and balanced training data, Random Forest revealed a performance of 77.0% accuracy and 85% areas under the curve, making it the most effective prediction model. The age at which family planning was first used, a partner or husband older than 40, being single, being Muslim, being pregnant, having previously been pregnant, needing to have more children, having a son or daughter relationship to the head of the household, and unmet needs for spacing and limiting were the top predictors of intention to use family planning, according to the SHAP analysis based on the random forest model. The research findings indicate that a range of personal and cultural factors may be taken into account when enacting health policies to enhance family planning intentions in Ethiopia. Therefore it\u0026rsquo;s highly recommended that the intention of family planning use and initiation of family planning provision should become a standard of service delivery to achieve the 2030 SDGs.\u003c/p\u003e","manuscriptTitle":"Explainable machine learning algorithm to identify predictors of intention to use family planning among reproductive-age women in Ethiopia: Evidence from the performance monitoring and accountability (PMA) survey 2021 dataset","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-01-25 16:14:25","doi":"10.21203/rs.3.rs-3848375/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"409059d2-f687-4325-9b98-e1f6d7d5c67b","owner":[],"postedDate":"January 25th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-10-22T10:08:07+00:00","versionOfRecord":[],"versionCreatedAt":"2024-01-25 16:14:25","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3848375","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3848375","identity":"rs-3848375","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}