Training and Testing Healthcare Models using Machine Learning Graphical User Interface Development Environment (ML-GUIDE) Software

Training and Testing Healthcare Models using Machine Learning Graphical User Interface Development Environment (ML-GUIDE) Software | 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 Article Training and Testing Healthcare Models using Machine Learning Graphical User Interface Development Environment (ML-GUIDE) Software Tejinder Singh, Vibhuti Jaswal, Vidushi Jaswal, Ashu Rastogi, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8255495/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 Background Machine learning (ML) is transforming healthcare research, but developing models can be complex and time-consuming. We created ML-GUIDE, a no-code tool with a graphical interface, to make ML model building faster and more accessible. Methods We developed the ML-GUIDE software using Python language and the Model-View-Controller architecture. ML-GUIDE provides 40 regression and 25 classification algorithms. We used publicly available Pima Indians Diabetes dataset to assess efficacy of the software. As per literature on ML models for Pima dataset, we developed, trained, and compared results from five diabetes prediction models: Adaboost, Decision Tree, Logistic Regression, Random Forest Classifier, and Support Vector Classifier for this study. We used metrics such as accuracy, average precision, and F1 score to evaluate the performances of ML models. Results We found that the model developed using ML-GUIDE software performed similar to published models. The variability (minimum to maximum) in accuracy of models in literature compared to models developed in this study ranges from 0.15% to 7.54%. The study also revealed that logistic regression with accuracy (78.65%), F1 score (0.849), and average precision (0.776) outperformed other models in predicting diabetes. Conclusion The results demonstrated that ML-GUIDE—a coding free tool matches performance of manual, time-consuming and expert driven algorithms. Therefore, ML-GUIDE tool can be used to rapidly test, develop and deploy ML-based healthcare solutions. Biological sciences/Computational biology and bioinformatics Health sciences/Health care Physical sciences/Mathematics and computing Health sciences/Medical research Figures Figure 1 Figure 2 Figure 3 1. Introduction In digital era, medical data is growing at an unprecedented scale—Machine Learning (ML) and Artificial Intelligence have become indispensable for making data informed decisions. There are many commercial ML & AI tools such as BigML 1 , Microsoft Azure ML studio 2 , Amazon SageMaker 3 , Google Cloud AutoML 4 , RapidMiner 5 , and DataRobot 6 are available to researchers. The high cost of purchasing, training, and data privacy, however, pose significant barriers to widespread adoption in healthcare. In recent years, there is a major shift to use language-based tools such as ChatGPT, Gemini, Claude, and Grok to generate ML codes. However, as project complexity increases, the effectiveness of these tools diminishes due to hallucinations, evolving landscape, and limited control 7 . Additionally, the integration of commercial tools in healthcare is affected by the limited funding 8 . It is more so for developing countries—paradoxically, these countries require more funds. Therefore, researchers from small research groups, institutes, and developing countries do not get the desired benefit of advances in ML & AI-based tools. The increasing speed, scale, and sharing of data pushed development of open-source solutions. We reviewed PubMed and identified several open-source ML software packages with graphical user interfaces (GUIs): ENNGene 9 , HDG-select 10 , ARX 11 , MetaboAnalyst 12 , FeAture Explorer (FAE) 13 , ARTMO's MLCA toolbox 14 , PDAUG 15 , and MVPANI 16 . Table 1 provides the detailed comparison of the open-source tools. Despite the promise, most tools are task-specific and do not cater to the broader needs of the healthcare research community. Operating a "Biostatistics Clinic" 17 at PGIMER-Chandigarh, we've observed an increasing demand for ML-based techniques to analyse data. Despite growing importance and realization of ML and AI techniques for enhancing patient care, the adoption of same in routine healthcare practice faces significant challenges 18 such as: (1) the need for advanced statistical knowledge, (2) programming proficiency, and (3) the rapidly evolving AI landscape. These challenges often resulted in ML-based healthcare systems that are complex, inefficient, time-consuming, and cost intensive 19 . To overcome these limitations, our key objective was to design, develop and deploy ML tools that require no coding and minimal training; similar to user-friendly software such as SPSS 20 and STATA 21 , ML-GUIDE allows researchers to experiment with ML models efficiently within their limited time. The objectives of this paper are as follows: first, to develop a no-code ML tool for researchers; second, to evaluate its effectiveness; and third, to facilitate deployment of trained models in practical, real-world settings. Machine Learning is not just about creating smart models; it's about deploying them to work in real life—publication to practice. However, an Industrial survey 22 estimated that only 20% models are deployed and the most significant hurdle (35% of the time) to do the same is lack of technical expertise. In line with this, we have provided a Python script file specifically designed to streamline the deployment process of models trained using ML-GUIDE software. Our goal is simple—from data to development to deployment of ML model so that machines can do the calculations and experts can concentrate on patient care without worrying about calculations. To access and validate the performance of ML-GUIDE software, we trained five ML algorithms namely—Adaboost. Decision Tree. Logistic Regression. Random Forest Classifier, and Support Vector Machine on PIMA Indians Diabetes dataset 23 using ML-GUIDE software and compared the results with published literature. This article is divided broadly into following sections: 1) Introduction to the tool, 2) Methodology for validating the tool, 3) Results, 4) Discussion, and 5) Conclusion. 2. Introduction to the Tool The ML-GUIDE (Machine Learning-Graphical User Interface Development Environment) software is an easy-to-use interface that simplifies the creation of ML models without requiring programming expertise; facilitating the development of customized ML models through a user-friendly point-and-click interface. We designed ML-GUIDE to provide diverse training and testing functionalities for ML models—researchers to focus on patient care rather than coding. ML-GUIDE has four main features: “ data preprocessing” , “ feature (variable) selection” , “ model building and training”, and “testing and evaluation” (definitions given in appendix-1). The model development process begins with data input in "CSV" format, where rows represent patients and columns represent features (variables). Figure 1 illustrates the ML-GUIDE homepage where the top row displays data preprocessing, model development, and training and testing windows. The output from the model is displayed on bottom left and dataset view is available on the bottom right. For data preprocessing the software has the following functionality: “ Handling Missing values” , “ Feature Encoding” (to convert categorical variables into numerical variables using One-hot encoding and Label encoding techniques), and “ Feature Scaling” (data normalisation and standardization). Feature selection is an essential step in the pipeline of ML model development and validation 24 . ML-GUIDE software provides three feature selection techniques—variance thresholding, SelectKBest, and recursive feature elimination algorithms for selecting the most informative features from the data to improve the performance and reduce the computational cost of training the ML model. Appendix 3 elucidates the differences between feature selection techniques. In the model-building phase, ML-GUIDE provides 25 classification algorithms and 40 regression algorithms from the scikit-learn 25 library in Python. For hyperparameter optimisation ML-GUIDE provides GridSearchCV algorithm which searches for the best combination of hyperparameters among the provided values for the given task. Appendix-4 enlists the ML algorithms available in the software. Trained models are saved as pickle files when the user opts to save models in the directory. For evaluation of trained models, ML-GUIDE provides a wide variety of evaluation metrics. For Binary Classification—AUC, F1 Score, Log Loss, Accuracy, and Average Precision. For Regression—Mean Squared Error, Mean Absolute Error, Explained Variance Score, Max Error, Mean Squared Log Error, and the R2 score. For Multiclass Classification the software generates a classification report containing Accuracy, Sensitivity, Specificity, and F1 Score. In addition to this, ML-GUIDE also provides visualization of results which is currently under development. 3. Materials and methods 3.1. Data We demonstrated and compared the performance of ML-GUIDE using the Pima Indians Diabetes dataset downloaded from Kaggle 23 . This dataset, originally from the National Institute of Diabetes and Digestive and Kidney Diseases, comprises health-related information from 768 (diabetic-268, non-diabetic-500) female patients of Pima Indian heritage, all aged at least 21 years. 3.1.1. Explanatory Variables Number of times pregnant, Plasma glucose concentration after a 2-hour oral glucose tolerance test, Diastolic blood pressure (mm Hg), Triceps skin fold thickness (mm), 2-hour serum insulin (mu U/ml), Body mass index, Diabetes pedigree function, and Age (years). 3.1.2. Outcome Variable Diabetes (present-1, absent-0) 3.2. Data Preprocessing The Pima Indian diabetes dataset did not contain any missing values, however, on closer inspection, we found noteworthy inconsistencies such as biologically implausible values (some values equal to Zero) in diastolic blood pressure (n = 235 (30.6%)) and blood glucose levels (n = 51 (6.7%)) variables. To address this, these instances were treated as missing values, allowing for more accurate data handling 26 . We imputed missing values with the respective median value for each attribute 27 . Furthermore, discrepancies were identified within the AGE (Inappropriate age values in some cases; <21 years) and Pregnancies (Unrealistic values such as 10 pregnancies at age 4); removing these anomalies enhanced the performance of the machine learning models in our study. We used normalisation for data uniformity—a crucial step for ML task; This process harmonised the ranges of each attribute, ensuring equal contributions to the machine learning algorithms 28 . 3.3. Feature Selection Data quality is the most significant contributor for the efficiency of machine learning algorithm 24 . We used Random Forest Classifier—Recursive Feature Elimination using five-fold cross-validation to select the most important features (features that contribute to output prediction) from the data 29 ; the process recursively remove the features with the least significance until performance on the validation folds starts decreasing. A vital advantage of this algorithm is that the users don’t have to specify desired number of features in advance. The algorithm automatically selects the set of features that results in highest performance during cross-validation. 3.4. Building and Training ML Models We compared five machine learning classifiers, namely: Adaboost, Decision Tree Classifier, Logistic Regression, Random Forest Classifier, and Support Vector Classifier using ML-GUIDE. We used these classifiers as the same were used and reported by the developer of ML models for Pima India Diabetes data. Modelling process is described in Fig. 2 . The ML models are described in appendix-8 for readers perusal. 3.5. Hyperparameter Optimization Hyperparameter optimisation is one of the most critical tasks in building efficient ML models 30 . We used GridSearchCV with 5-fold cross-validation to find the best combination of hyperparameters; currently it is the only option available in the software for hyperparameter optimisation. Using the ML-GUIDE software, we selected the best models from GridSearchCV for final training. 3.6. Testing and Evaluation We used “train_test_split” (train:test ratio 75:25) function in the software to partition Pima data into training and testing data. The test dataset, comprising a total of 192 participants (including diabetic and non-diabetic) were used to assess the efficacy of the trained models. The diabetic:non-diabetic participant ratio is same as original dataset. The models were evaluated using Accuracy, Average Precision, and F1 score. These metrics were instrumental in quantitatively gauging the models' predictive prowess; all these metrics are calculated based on True Positives (TP), False Positives (FP), True Negatives (TN), and False Negatives (FN). A \(\:\text{c}\text{c}\text{u}\text{r}\text{a}\text{c}\text{y}=\:\frac{\text{T}\text{P}+\text{T}\text{N}}{\text{T}\text{P}+\text{F}\text{P}+\text{T}\text{N}+\text{F}\text{N}}\) , F1 score - It is the harmonic mean of the recall (sensitivity) and precision (PPV), \(\:\:\text{F}1\:\text{s}\text{c}\text{o}\text{r}\text{e}=2\text{*}\frac{\text{P}\text{r}\text{e}\text{c}\text{i}\text{s}\text{i}\text{o}\text{n}\text{*}\text{R}\text{e}\text{c}\text{a}\text{l}\text{l}}{\text{P}\text{r}\text{e}\text{c}\text{i}\text{s}\text{i}\text{o}\text{n}+\:\text{R}\text{e}\text{c}\text{a}\text{l}\text{l}}\) . Figure-2 demonstrates the process from data selection to model evaluation used by authors. 3.7. Model Deployment The ML-GUIDE software saves the trained models as “.pkl” files (example: “AdaBoostClassifier.pkl”). To use trained models, researchers can copy the python script from supplementary file (Appendix-5)—paste this code into a new empty text file and save the file with the name “model.py” in the same folder where trained models are saved. After saving the “model.py” file, follow the steps: Install the Anaconda package. After installation, open the "Anaconda Prompt". To install the Streamlit app, type "pip install streamlit" in "Anaconda Prompt". Go to anaconda prompt and navigate to "model.py" directory using the “cd” command. For example, “cd models” to navigate to the “models” folder. The command "streamlit run model.py" in "Anaconda Prompt" will launch the app in default web browser. Specify the model path, for example, "E:/models/AdaBoostClassifier.pkl". To get the model path, right click on the model.pkl file and click properties. The path is given in the “Location” field. Copy the path and add model name following the forward slash (/), for example, “/AdaBoostClassifier.pkl” Follow the on-screen instructions, ensuring to include all the data preprocessing steps applied during model training. Following these steps the researchers may use the trained models in their computer for making predictions. The users may also watch the video provided with the manuscript for the same. 3.8 Hardware used in this study– We used a Windows 11 machine with Intel (R) Core ™ i5-8250U CPU @ 1.60GHz CPU and 16 GB RAM for testing of software. However, we have also tested the software on Windows 10 machine with Intel (R) Core ™ i5-8250U CPU @ 1.60GHz CPU and 8 GB RAM. Currently the software runs on Windows 10 and Windows 11 operating systems only. 4. Results We assessed the predictive performance of five ML algorithms developed and trained using the ML-GUIDE software for binary classification tasks—diabetes prediction. To evaluate the performance of ML-GUIDE, we reported the hyperparameter optimisation time, training time and model evaluation metrics (accuracy, F1 score and average precision) in Table 2 . Notably, in the context of Pima Indians Diabetes dataset, we observed that decision tree and logistic regression algorithms exhibited the shortest times for hyperparameter optimisation and less than 1 second for training of all models. The Random Forest algorithm took the longest time—242 seconds (almost double—4-minute) for hyperparameter optimisation when compared to Adaboost (129 seconds—2 minute). It's important to note that the time taken by ML models is influenced by various factors, including the model's complexity, dataset size, and the processing power of the machine 31 . Table 1 Features of free open source no-code ML tools compared to the ML-GUIDE software S. no. Tool Number of algorithms provided in the tool Input type Data pre-processing Feature Selection/ Extraction Data Visualisation Classification Regression 1 ML-GUIDE 25 40 “CSV” √ √ √ 2 ENNGene 1 - BED genomic interval files √ X √ 3 HDG-select 1 - GEO “SOFT” and “CSV” √ √ X 4 ARX 3 - “CSV” √ √ √ 5 MetaboAnalyst 3 - Wide variety of formats for metabolomics data √ √ √ 6 FeAture Explorer (FAE) 10 - “CSV” √ √ √ 7 ARTMO's MLCA toolbox 19 - “txt” X √ √ 8 PDAUG 8 - Peptide data √ √ √ 9 MVPANI 7 4 “txt”, “xlsx”, and “mat” √ √ X Table 2 Performance evaluation of ML-GUIDE software for developing ML models on the PIMA Indians Diabetes dataset S. no. ML Model Hyperparameter optimisation time (seconds) Training time (seconds) Accuracy (%) F1 score Average Precision 1 Adaboost 129 < 1 75.52 0.83 0.74 2 Decision Tree 2 < 1 71.35 0.79 0.74 3 Logistic Regression 3 < 1 78.65 0.85 0.78 4 Random Forest 242 < 1 77.60 0.83 0.79 5 Support Vector Machine 5 < 1 77.08 0.84 0.77 Logistic Regression emerged as the best performer among the evaluated models, demonstrating top performance across multiple key metrics. Logistic Regression achieved the highest scores in terms of Accuracy and F1 Score. Random Forest Classifier outperformed all the other models on Average Precision metrics. Equally noteworthy are the competitive performances exhibited by the AdaBoost Classifier and the Support Vector Machine. Decision Tree was the worst performer among all the models. Figure 3 represents the comparison of ML models—automated coding off all the models in the current study with manually developed models on PIMA Indians diabetes dataset 27 , 32 – 40 . Please refer appendix-6 for table representing accuracies of different models. The results demonstrate promising performance across multiple models compared to published results. Notably, the AdaBoost Classifier achieved an accuracy of 75.52%, closely approaching the literature-reported accuracy of 78.3%. The DecisionTree Classifier also displayed competitive results, with an accuracy of 71.35%, surpassing some of the reported literature values. Logistic Regression and the Support Vector Machine algorithms outperformed the published results with an accuracy of 78.65% and 77.08% respectively, while the Random Forest Classifier achieved an accuracy of 77.60%, maintaining competitive performance. 5. Discussion We developed the ML-GUIDE software for medical researchers to expedite development, training, and testing of machine learning models by eliminating manual algorithm coding process. The surge in medical data prompts researchers to use machine learning for improved patient care 41 , 42 . The coding prerequisites, patient load, and time and resource constraints, however, hinder health experts from experimenting with various ML techniques 19 . To mitigate these challenges, we had three objectives, first, to develop the no-code ML tool for researchers, second, to test the efficacy of this tool, and third, to facilitate researchers in deploying their trained models into practice. There are multiple advantages of using ML-GUIDE tool compared to slow, manual, and expert driven algorithmic process. The tool offers a user-friendly Graphical User Interface (GUI), enabling researchers to experiment with machine learning (ML) algorithms without learning to code. Unlike the limited diversity of ML models in open-source no-code tools available on PubMed 9 – 16 , 43 , 44 , the ML-GUIDE software provides a comprehensive selection of ML algorithms, encompassing 25 classification and 40 regression algorithms. Additionally, it allows users to manually adjust key hyperparameters or utilize hyperparameter optimization techniques to fine-tune model performance. Ensuring the reliability and generalizability of trained ML models requires testing on external datasets. The ML-GUIDE software simplifies this process with a one-click option for testing trained models on external datasets. Notably, the increasing use of ML techniques in healthcare mirrors the earlier adoption of statistical software like SPSS and STATA. Just as these platforms made advanced statistical methods accessible to non-statisticians, ML-GUIDE aims to provide access to state-of-the-art ML algorithms to non-coders in clinical and research settings. To assess the software's effectiveness, we developed five ML models using ML-GUIDE and compared their accuracy with published literature 27 , 32 – 40 . We used Pima Indians Diabetes dataset to train these models. Figure 3 illustrates that the predictive accuracy of ML models generated with ML-GUIDE aligns with those published in the literature. There are minor variations in the accuracy of diabetes prediction among models from different studies. The variability (minimum to maximum) in accuracy of models in literature compared to models developed in this study ranges from 0.15% to 7.54%; this can be attributed to variations in data preprocessing, feature selection, and hyperparameter optimisation techniques used in different studies 45 – 47 . Our results confirm that the ML models created using the ML-GUIDE software perform at par with the models found in existing literature. Moreover, tests on the Pima Indians diabetes dataset also showed that each model (depending upon processor speed) can be trained in less than one second using 576 samples. The limited sample size does not strongly emphasize the speed of model training in the software, but it does indicate that ML-GUIDE operates within a reasonable timeframe. Any ML model regardless of their training is only useful when deployed into practice; the crucial test is whether they improve patient outcomes or not 48 . We couldn't find any software articles in our review 9 – 16 , 43 , 44 that helped researchers deploy their machine learning models. To deploy and use the trained models the ML-GUIDE software enables users to save the trained models as “’pkl” files and also save a "txt" file documenting the model development steps (pipeline) for future reference. We provided a Python script as guidance for researchers to deploy their trained ML models to test them in real-time clinical settings. This code can also be utilised for deploying models on an online server to build web applications. However, we have not provided the procedure to integrate trained machine learning models into web applications as it is beyond the scope of the current article. One of the noteworthy aspects of the ML-GUIDE tool is its ability to make machine learning more accessible, especially for non-programmers. This inclusiveness allows a broader community of healthcare researchers to engage in machine learning exploration for addressing complex healthcare challenges. We encourage the researchers to extensively explore the software's capabilities and test its limitations. 6. Conclusion Many researchers intend to use machine learning in healthcare, but lack of time, programming expertise, and resources slows them down. To facilitate healthcare researchers, we created ML-GUIDE—a simple tool that lets users quickly build ML models with a point-and-click interface. By eliminating the need for programming: a significant barrier for non-programmers, ML-GUIDE empowers researchers to engage in development of ML-based applications. This open-access tool offers a wide range of algorithms, data preprocessing methods, and hyperparameter optimization options. Our evaluation shows that ML-GUIDE produces models with predictive accuracy on par with those found in existing literature. Declarations Ethics declaration: Not applicable . Competing interests All authors declare no financial or non-financial competing interests. Funding None Data Sharing The study is conducted on the publicly available Pima Indians Diabetes Dataset, which can be downloaded from the Kaggle website: https://www.kaggle.com/datasets/uciml/pima-indians-diabetes-database . References BigML. https://bigml.com/ D Mukunthu & S Gillett. Announcing automated ML capability in azure machine learning. (2018). https://azure.microsoft.com/en-us/blog/announcing-automated-ml-capability-in-azure-machine-learning/ Amazon SageMaker. (2019). https://aws.amazon.com/sagemaker/ Cloud, A. M. L. Cloud AutoML (2019). https://cloud.google.com/automl/ RapidMiner https://rapidminer.com/ DataRobot Automated Machine Learning. (2019). https://www.datarobot.com/platform/automated-machine-learning/ Zhang, Z., Wang, Y., Wang, C., Chen, J. & Zheng, Z. LLM Hallucinations in Practical Code Generation: Phenomena, Mechanism, and Mitigation. Preprint at (2025). https://doi.org/10.48550/arXiv.2409.20550 Alami, H. et al. Understanding the integration of artificial intelligence in healthcare organisations and systems through the NASSS framework: a qualitative study in a leading Canadian academic centre. BMC Health Serv. Res. 24 , 701 (2024). Chalupová, E. et al. ENNGene: an Easy Neural Network model building tool for Genomics. BMC Genom. 23 , 248 (2022). Hameed, S. S., Hassan, R., Hassan, W. H., Muhammadsharif, F. F. & Latiff, L. A. HDG-select: A novel GUI based application for gene selection and classification in high dimensional datasets. PloS One . 16 , e0246039 (2021). Eicher, J., Bild, R., Spengler, H., Kuhn, K. A. & Prasser, F. A comprehensive tool for creating and evaluating privacy-preserving biomedical prediction models. BMC Med. Inf. Decis. Mak. 20 , 29 (2020). Xia, J., Psychogios, N., Young, N. & Wishart, D. S. MetaboAnalyst: a web server for metabolomic data analysis and interpretation. Nucleic Acids Res. 37 , W652–660 (2009). Song, Y. et al. FeAture Explorer (FAE): A tool for developing and comparing radiomics models. PloS One . 15 , e0237587 (2020). Aghababaei, M. et al. Introducing ARTMO’s Machine-Learning Classification Algorithms Toolbox: Application to Plant-Type Detection in a Semi-Steppe Iranian Landscape. Remote Sens. 14 , 4452 (2022). Joshi, J. & Blankenberg, D. PDAUG: a Galaxy based toolset for peptide library analysis, visualization, and machine learning modeling. BMC Bioinform. 23 , 197 (2022). Peng, Y. et al. MVPANI: A Toolkit With Friendly Graphical User Interface for Multivariate Pattern Analysis of Neuroimaging Data. Front. Neurosci. 14 , 545 (2020). Kamal, K., Sharma, M., Rahul, M., Rakesh, K. & Koushal, V. Ensconcing a biostatistics clinic in tertiary care research institute of India: A descriptive study. J. Fam Med. Prim. Care . 11 , 2051 (2022). Korot, E. et al. Code-free deep learning for multi-modality medical image classification. Nat. Mach. Intell. 3 , 288–298 (2021). Waring, J., Lindvall, C. & Umeton, R. Automated machine learning: Review of the state-of-the-art and opportunities for healthcare. Artif. Intell. Med. 104 , 101822 (2020). IBM. IBM SPSS. STATA. Eric Siegel. Models Are Rarely Deployed: An Industry-wide Failure in Machine Learning Leadership. https://www.kdnuggets.com/2022/01/models-rarely-deployed-industrywide-failure-machine-learning-leadership.html Pima Indians Diabetes Database - Kaggle. Domingos, P. A few useful things to know about machine learning. Commun. ACM . 55 , 78–87 (2012). Pedregosa, F. et al. Scikit-learn: Machine Learning in Python. Preprint at (2018). http://arxiv.org/abs/1201.0490 Pima Indians Diabetes. https://data.world/uci/pima-indians-diabetes Kumari, S., Kumar, D. & Mittal, M. An ensemble approach for classification and prediction of diabetes mellitus using soft voting classifier. Int. J. Cogn. Comput. Eng. 2 , 40–46 (2021). Li, W. & Liu, Z. A method of SVM with Normalization in Intrusion Detection. Procedia Environ. Sci. 11 , 256–262 (2011). Darst, B. F., Malecki, K. C. & Engelman, C. D. Using recursive feature elimination in random forest to account for correlated variables in high dimensional data. BMC Genet. 19 , 65 (2018). Frank & Hutter Holger Hoos, & Kevin Leyton-Brown. An efficient approach for assessing hyperparameter importance. (2014). L’Heureux, A., Grolinger, K., Elyamany, H. F. & Capretz, M. A. M. Machine Learning With Big Data: Challenges and Approaches. IEEE Access. 5 , 7776–7797 (2017). Dhilsath Fathima, M. & Justin Samuel, S. Improved Adaboost Algorithm with Regression Imputation for Prediction of Chronic Type 2 Diabetes Mellitus. in Communication and Intelligent Systems (eds Sharma, H., Gupta, M. K., Tomar, G. S. & Lipo, W.) (, vol 204 691–708 (Springer Singapore, Singapore, (2021). Bozkurt, M. R., Yurtay, N., Yilmaz, Z. & Sertkaya, C. Comparison of different methods for determining diabetes. Turk. J. Electr. Eng. Comput. Sci. 22 , 1044–1055 (2014). Chatrati, S. P. et al. Smart home health monitoring system for predicting type 2 diabetes and hypertension. J. King Saud Univ. - Comput. Inf. Sci. 34 , 862–870 (2022). Deng, D. & Kasabov, N. On-line pattern analysis by evolving self-organizing maps. Neurocomputing 51 , 87–103 (2003). Chang, V., Bailey, J., Xu, Q. A. & Sun, Z. Pima Indians diabetes mellitus classification based on machine learning (ML) algorithms. Neural Comput. Appl. 35 , 16157–16173 (2023). Rajendra, P. & Latifi, S. Prediction of diabetes using logistic regression and ensemble techniques. Comput. Methods Programs Biomed. Update . 1 , 100032 (2021). Saxena, R., Sharma, S. K., Gupta, M. & Sampada, G. C. A Novel Approach for Feature Selection and Classification of Diabetes Mellitus: Machine Learning Methods. Comput. Intell. Neurosci. 1–11 (2022). (2022). Zou, Q. et al. Predicting Diabetes Mellitus With Machine Learning Techniques. Front. Genet. 9 , 515 (2018). Karthikeyani, V., Parvin Begum, I. & Tajudin, K. Shahina Begam, I. Comparative of Data Mining Classification Algorithm (CDMCA) in Diabetes Disease Prediction. Int. J. Comput. Appl. 60 , 26–31 (2012). Baro, E., Degoul, S., Beuscart, R. & Chazard, E. Toward a Literature-Driven Definition of Big Data in Healthcare. BioMed Res. Int. 1–9 (2015). (2015). Hay, S. I., George, D. B., Moyes, C. L. & Brownstein, J. S. Big data opportunities for global infectious disease surveillance. PLoS Med. 10 , e1001413 (2013). Razeghi, O. et al. CemrgApp: An interactive medical imaging application with image processing, computer vision, and machine learning toolkits for cardiovascular research. SoftwareX 12 , 100570 (2020). Mattonen, S. A. et al. Quantitative imaging feature pipeline: a web-based tool for utilizing, sharing, and building image-processing pipelines. J. Med. Imaging Bellingham Wash. 7 , 042803 (2020). Ahsan, M., Mahmud, M., Saha, P., Gupta, K. & Siddique, Z. Effect of Data Scaling Methods on Machine Learning Algorithms and Model Performance. Technologies 9 , 52 (2021). Ejiro Onose. How to Solve Reproducibility in ML. https://neptune.ai/blog/how-to-solve-reproducibility-in-ml Ghorbani, R. & Ghousi, R. Comparing Different Resampling Methods in Predicting Students’ Performance Using Machine Learning Techniques. IEEE Access. 8 , 67899–67911 (2020). Kelly, C. J., Karthikesalingam, A., Suleyman, M., Corrado, G. & King, D. Key challenges for delivering clinical impact with artificial intelligence. BMC Med. 17 , 195 (2019). Additional Declarations No competing interests reported. Supplementary Files SupplementryFile.docx ModelDeployment.mp4 MLGuide.mp4 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-8255495","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":558905831,"identity":"52c39f09-f7c1-400b-ae5e-230adc27b23e","order_by":0,"name":"Tejinder Singh","email":"","orcid":"","institution":"Post Graduate Institute of Medical Education and Research","correspondingAuthor":false,"prefix":"","firstName":"Tejinder","middleName":"","lastName":"Singh","suffix":""},{"id":558905834,"identity":"1c2e3a38-6bbe-4f4e-bdd5-c18e2777c8dc","order_by":1,"name":"Vibhuti Jaswal","email":"","orcid":"","institution":"Army Institute of Law","correspondingAuthor":false,"prefix":"","firstName":"Vibhuti","middleName":"","lastName":"Jaswal","suffix":""},{"id":558905837,"identity":"c61746ec-8f46-46dc-b797-1d3599523bc9","order_by":2,"name":"Vidushi Jaswal","email":"","orcid":"","institution":"Mehr Chand Mahajan DAV College for Womeni","correspondingAuthor":false,"prefix":"","firstName":"Vidushi","middleName":"","lastName":"Jaswal","suffix":""},{"id":558905839,"identity":"7ba374cf-fb2d-420b-8f70-bfa4889e3816","order_by":3,"name":"Ashu Rastogi","email":"","orcid":"","institution":"Post Graduate Institute of Medical Education and Research","correspondingAuthor":false,"prefix":"","firstName":"Ashu","middleName":"","lastName":"Rastogi","suffix":""},{"id":558905841,"identity":"ccd49a04-7c3b-4341-9906-eacabdc2a61a","order_by":4,"name":"Kamal Kishore","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+UlEQVRIiWNgGAWjYBACCR7GBmYY58AHIMHGToqWgzNAWpjxKIdoYWCAq2HmAZMEtEj2HG7+XFBxT45Buv3iYZtf2+T5mBkYP3zMwa1FmrexTXrGmWJjBpkzBYdz+24btjEzMEvO3IZbixw/Yxszb1tCYoNETsLh3J7bQC7QO7z4tTR/BmqpB2ux7LltT1AL0GEN0kAtCQwS6QcOM/y4nUhQi2TPwTZpnjMJhm0yZxgO9jbcTm5jZmzG6xeJM+mPP/NUJMjzS7c//vDjz23b+e3NBz98xKMFDtgkeAwYGNtATMYGItSD7WN/wMDwh0jFo2AUjIJRMKIAAO85TdXI3mG7AAAAAElFTkSuQmCC","orcid":"","institution":"Post Graduate Institute of Medical Education and Research","correspondingAuthor":true,"prefix":"","firstName":"Kamal","middleName":"","lastName":"Kishore","suffix":""}],"badges":[],"createdAt":"2025-12-02 03:38:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8255495/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8255495/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":98079307,"identity":"b5286f46-87c0-4abc-9231-b34aaeb1ca69","added_by":"auto","created_at":"2025-12-12 14:29:41","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":355100,"visible":true,"origin":"","legend":"","description":"","filename":"MLGUIDEPaperScientificReportsupdated.docx","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/caf520fc849b7e5a8f5509a6.docx"},{"id":98429207,"identity":"1299b3ba-9947-4ea5-b405-0b87a4617546","added_by":"auto","created_at":"2025-12-17 16:42:57","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6811,"visible":true,"origin":"","legend":"","description":"","filename":"10c96f65c9c248758e0e809268041569.json","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/79a35942f3ac8815443ec5d5.json"},{"id":98079320,"identity":"b91f5058-08e1-4495-83e1-8b4423252a5e","added_by":"auto","created_at":"2025-12-12 14:29:42","extension":"mp4","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":25546492,"visible":true,"origin":"","legend":"","description":"","filename":"MLGuide.mp4","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/5b77782283b70f402d78a904.mp4"},{"id":98429592,"identity":"14b1902a-e7c3-40d4-8462-8faea8e085f3","added_by":"auto","created_at":"2025-12-17 16:43:48","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":78929,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementryFile.docx","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/8e32e11238c507fa39bb6e48.docx"},{"id":98430020,"identity":"5315cf8b-87e5-4636-94ce-d7db63eba7e4","added_by":"auto","created_at":"2025-12-17 16:44:37","extension":"xml","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":107601,"visible":true,"origin":"","legend":"","description":"","filename":"10c96f65c9c248758e0e8092680415691enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/0a17df4dce4236417eb44411.xml"},{"id":98430026,"identity":"95ca140b-7a33-4126-b6a0-c98a3b50070f","added_by":"auto","created_at":"2025-12-17 16:44:39","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":148480,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/435d5321b3fda143b89dc66d.png"},{"id":98079314,"identity":"3dd33f8e-cb5e-47b0-95ea-7b28db7f078c","added_by":"auto","created_at":"2025-12-12 14:29:42","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1015267,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/1a38d6577e13a355a3bd3e6c.jpeg"},{"id":98429600,"identity":"21b875cb-0ae3-4405-b2ab-8705872435a1","added_by":"auto","created_at":"2025-12-17 16:43:49","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":76861,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/9b534258f8f36323fdc51abc.png"},{"id":98079310,"identity":"fd58a3e4-f88d-4817-9e41-a3a836db4a5f","added_by":"auto","created_at":"2025-12-12 14:29:42","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":31869,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/93bfe4e88c2b1df480f47093.png"},{"id":98079318,"identity":"fc390aed-a7f5-44bd-beb2-2db9aa1a9ccb","added_by":"auto","created_at":"2025-12-12 14:29:42","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":168318,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/da029513ee895e174a624075.png"},{"id":98079315,"identity":"91917447-222d-4cdf-a54e-3fd15cf5514a","added_by":"auto","created_at":"2025-12-12 14:29:42","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":21723,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/492981ce04fbbfcfa76b1720.png"},{"id":98079319,"identity":"ae30a818-5997-4943-ada7-eff1eb623108","added_by":"auto","created_at":"2025-12-12 14:29:42","extension":"xml","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102449,"visible":true,"origin":"","legend":"","description":"","filename":"10c96f65c9c248758e0e8092680415691structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/1cf4365683d5314716760187.xml"},{"id":98079317,"identity":"bb1f140d-dd88-4906-acf8-2f899386d78d","added_by":"auto","created_at":"2025-12-12 14:29:42","extension":"html","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":119772,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/c5da72c7bf46fe38b0a915a8.html"},{"id":98079304,"identity":"e86c7b93-ead8-4f2c-be96-0baa806bc143","added_by":"auto","created_at":"2025-12-12 14:29:41","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":261866,"visible":true,"origin":"","legend":"\u003cp\u003eThe snippet of ML-GUIDE software homepage\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/273f8273faff32265df742b9.png"},{"id":98429307,"identity":"d992cadb-67ad-46b0-b214-9a576b106fa6","added_by":"auto","created_at":"2025-12-17 16:43:11","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":928045,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart depicting the steps followed in developing ML models for second objective. The ML-GUIDE software provides point-and-click based interface to perform all actions without coding.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/fe8fb8728439a5e6d988af00.png"},{"id":98079305,"identity":"24be04d6-7cfc-4aeb-8048-8d3e6772a887","added_by":"auto","created_at":"2025-12-12 14:29:41","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":182369,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of classification accuracy of ML models developed in this study on Pima Indians Diabetes dataset to different machine learning methods published in the literature.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/d07a2ad3c37a36b9cb8a56de.png"},{"id":98444707,"identity":"08bebf2b-7fc5-40a8-9328-8014344d8ce9","added_by":"auto","created_at":"2025-12-17 17:17:01","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2186664,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/fafb7b1c-4b73-463f-bf29-ef01f0aac7d3.pdf"},{"id":98079309,"identity":"632bb3cb-94cb-441b-b66a-e09fed51a061","added_by":"auto","created_at":"2025-12-12 14:29:42","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":78929,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementryFile.docx","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/789b294149b3e99c0aaeafe0.docx"},{"id":98079321,"identity":"b8fc27bd-c77b-4e35-9656-89bae19bb98b","added_by":"auto","created_at":"2025-12-12 14:29:42","extension":"mp4","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":11635443,"visible":true,"origin":"","legend":"","description":"","filename":"ModelDeployment.mp4","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/a6ae42630ac78e8bbdae0274.mp4"},{"id":98079323,"identity":"2ceb9290-9811-4573-896a-e904442690ab","added_by":"auto","created_at":"2025-12-12 14:29:42","extension":"mp4","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":25546492,"visible":true,"origin":"","legend":"","description":"","filename":"MLGuide.mp4","url":"https://assets-eu.researchsquare.com/files/rs-8255495/v1/bafea4b51ca23cbe7a0c6c36.mp4"}],"financialInterests":"No competing interests reported.","formattedTitle":"Training and Testing Healthcare Models using Machine Learning Graphical User Interface Development Environment (ML-GUIDE) Software","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eIn digital era, medical data is growing at an unprecedented scale\u0026mdash;Machine Learning (ML) and Artificial Intelligence have become indispensable for making data informed decisions.\u0026nbsp;There are many commercial ML \u0026amp; AI tools such as BigML\u003csup\u003e1\u003c/sup\u003e, Microsoft Azure ML studio\u003csup\u003e2\u003c/sup\u003e, Amazon SageMaker\u003csup\u003e3\u003c/sup\u003e, Google Cloud AutoML\u003csup\u003e4\u003c/sup\u003e, RapidMiner\u003csup\u003e5\u003c/sup\u003e, and DataRobot\u003csup\u003e6\u003c/sup\u003e are available to researchers. The high cost of purchasing, training, and data privacy, however, pose significant barriers to widespread adoption in healthcare.\u0026nbsp;In recent years, there is a major shift to use language-based tools such as ChatGPT, Gemini, Claude, and Grok to generate ML codes. However, as project complexity increases, the effectiveness of these tools diminishes due to hallucinations, evolving landscape, and limited control\u003csup\u003e7\u003c/sup\u003e.\u0026nbsp;Additionally, the integration of commercial tools in healthcare is affected by the limited funding\u003csup\u003e8\u003c/sup\u003e. It is more so for developing countries\u0026mdash;paradoxically, these countries require more funds. Therefore, researchers from small research groups, institutes, and developing countries do not get the desired benefit of advances in ML \u0026amp; AI-based tools.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe increasing speed, scale, and sharing of data pushed development of\u0026nbsp;open-source solutions. We reviewed PubMed and identified several open-source ML software packages with graphical user interfaces (GUIs):\u0026nbsp;ENNGene\u003csup\u003e9\u003c/sup\u003e, HDG-select\u003csup\u003e10\u003c/sup\u003e,\u0026nbsp;ARX\u003csup\u003e11\u003c/sup\u003e, MetaboAnalyst\u003csup\u003e12\u003c/sup\u003e, FeAture Explorer (FAE)\u003csup\u003e13\u003c/sup\u003e,\u0026nbsp;ARTMO\u0026apos;s MLCA toolbox\u003csup\u003e14\u003c/sup\u003e, PDAUG\u003csup\u003e15\u003c/sup\u003e, and MVPANI\u003csup\u003e16\u003c/sup\u003e. Table 1 provides the detailed comparison of the open-source tools. Despite the promise, most tools are task-specific and do not cater to the broader needs of the healthcare research community. Operating a \u0026quot;Biostatistics Clinic\u0026quot;\u003csup\u003e17\u003c/sup\u003e at PGIMER-Chandigarh, we\u0026apos;ve observed an increasing demand for ML-based techniques to analyse data.\u0026nbsp;Despite growing importance and realization of ML and AI techniques for enhancing patient care, the adoption of same in routine healthcare practice faces significant challenges\u003csup\u003e18\u003c/sup\u003e such as: (1) the need for advanced statistical knowledge, (2) programming proficiency, and (3) the rapidly evolving AI landscape. These challenges often resulted in ML-based healthcare systems that are complex, inefficient, time-consuming, and cost intensive\u003csup\u003e19\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo overcome these limitations, our key objective was\u0026nbsp;to design, develop and deploy ML tools that require no coding and minimal training; similar to user-friendly software such as SPSS\u003csup\u003e20\u003c/sup\u003e and STATA\u003csup\u003e21\u003c/sup\u003e, ML-GUIDE allows researchers to experiment with ML models efficiently within their limited time. The objectives of this paper are as follows: first, to develop a no-code ML tool for researchers; second, to evaluate its effectiveness; and third, to facilitate deployment of trained models in practical, real-world settings.\u003c/p\u003e\n\u003cp\u003eMachine Learning is not just about creating smart models; it\u0026apos;s about deploying them to work in real life\u0026mdash;publication to practice. However, an Industrial survey\u003csup\u003e22\u003c/sup\u003e estimated that only 20% models are deployed and the most significant hurdle (35% of the time) to do the same is lack of technical expertise. In line with this, we have provided a Python script file specifically designed to streamline the deployment process of models trained using ML-GUIDE software. Our goal is simple\u0026mdash;from data to development to deployment of ML model so that machines can do the calculations and experts can concentrate on patient care without worrying about calculations.\u003c/p\u003e\n\u003cp\u003eTo access and validate the performance of ML-GUIDE software, we trained five ML algorithms namely\u0026mdash;Adaboost. Decision Tree. Logistic Regression. Random Forest Classifier, and Support Vector Machine on PIMA Indians Diabetes dataset\u003csup\u003e23\u003c/sup\u003e using ML-GUIDE software and compared the results with published literature. This article is divided broadly into following sections: 1) Introduction to the tool, 2) Methodology for validating the tool, 3) Results, 4) Discussion, and 5) Conclusion.\u003c/p\u003e"},{"header":"2. Introduction to the Tool","content":"\u003cp\u003eThe ML-GUIDE (Machine Learning-Graphical User Interface Development Environment) software is an easy-to-use interface that simplifies the creation of ML models without requiring programming expertise; facilitating the development of customized ML models through a user-friendly point-and-click interface. We designed ML-GUIDE to provide diverse training and testing functionalities for ML models\u0026mdash;researchers to focus on patient care rather than coding. ML-GUIDE has four main features: \u0026ldquo;\u003cem\u003edata preprocessing\u0026rdquo;\u003c/em\u003e, \u0026ldquo;\u003cem\u003efeature (variable) selection\u0026rdquo;\u003c/em\u003e, \u0026ldquo;\u003cem\u003emodel building and training\u0026rdquo;, and \u0026ldquo;testing and evaluation\u0026rdquo;\u003c/em\u003e (definitions given in appendix-1). The model development process begins with data input in \u003cem\u003e\"CSV\"\u003c/em\u003e format, where rows represent patients and columns represent features (variables). Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e illustrates the ML-GUIDE homepage where the top row displays data preprocessing, model development, and training and testing windows. The output from the model is displayed on bottom left and dataset view is available on the bottom right.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eFor data preprocessing the software has the following functionality: \u0026ldquo;\u003cem\u003eHandling Missing values\u0026rdquo;\u003c/em\u003e, \u0026ldquo;\u003cem\u003eFeature Encoding\u0026rdquo;\u003c/em\u003e (to convert categorical variables into numerical variables using One-hot encoding and Label encoding techniques), and \u0026ldquo;\u003cem\u003eFeature Scaling\u0026rdquo;\u003c/em\u003e (data normalisation and standardization). Feature selection is an essential step in the pipeline of ML model development and validation\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. ML-GUIDE software provides three feature selection techniques\u0026mdash;variance thresholding, SelectKBest, and recursive feature elimination algorithms for selecting the most informative features from the data to improve the performance and reduce the computational cost of training the ML model. Appendix 3 elucidates the differences between feature selection techniques.\u003c/p\u003e\u003cp\u003eIn the model-building phase, ML-GUIDE provides 25 classification algorithms and 40 regression algorithms from the scikit-learn\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e library in Python. For hyperparameter optimisation ML-GUIDE provides GridSearchCV algorithm which searches for the best combination of hyperparameters among the provided values for the given task. Appendix-4 enlists the ML algorithms available in the software. Trained models are saved as pickle files when the user opts to save models in the directory. For evaluation of trained models, ML-GUIDE provides a wide variety of evaluation metrics. For Binary Classification\u0026mdash;AUC, F1 Score, Log Loss, Accuracy, and Average Precision. For Regression\u0026mdash;Mean Squared Error, Mean Absolute Error, Explained Variance Score, Max Error, Mean Squared Log Error, and the R2 score. For Multiclass Classification the software generates a classification report containing Accuracy, Sensitivity, Specificity, and F1 Score. In addition to this, ML-GUIDE also provides visualization of results which is currently under development.\u003c/p\u003e"},{"header":"3. Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e3.1. Data\u003c/h2\u003e\u003cp\u003eWe demonstrated and compared the performance of ML-GUIDE using the Pima Indians Diabetes dataset downloaded from Kaggle\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. This dataset, originally from the National Institute of Diabetes and Digestive and Kidney Diseases, comprises health-related information from 768 (diabetic-268, non-diabetic-500) female patients of Pima Indian heritage, all aged at least 21 years.\u003c/p\u003e\u003cdiv id=\"Sec4\" class=\"Section3\"\u003e\u003ch2\u003e3.1.1. Explanatory Variables\u003c/h2\u003e\u003cp\u003eNumber of times pregnant, Plasma glucose concentration after a 2-hour oral glucose tolerance test, Diastolic blood pressure (mm Hg), Triceps skin fold thickness (mm), 2-hour serum insulin (mu U/ml), Body mass index, Diabetes pedigree function, and Age (years).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section3\"\u003e\u003ch2\u003e3.1.2. Outcome Variable\u003c/h2\u003e\u003cp\u003eDiabetes (present-1, absent-0)\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e3.2. Data Preprocessing\u003c/h2\u003e\u003cp\u003eThe Pima Indian diabetes dataset did not contain any missing values, however, on closer inspection, we found noteworthy inconsistencies such as biologically implausible values (some values equal to Zero) in diastolic blood pressure (n\u0026thinsp;=\u0026thinsp;235 (30.6%)) and blood glucose levels (n\u0026thinsp;=\u0026thinsp;51 (6.7%)) variables. To address this, these instances were treated as missing values, allowing for more accurate data handling\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. We imputed missing values with the respective median value for each attribute\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. Furthermore, discrepancies were identified within the AGE (Inappropriate age values in some cases; \u0026lt;21 years) and Pregnancies (Unrealistic values such as 10 pregnancies at age 4); removing these anomalies enhanced the performance of the machine learning models in our study. We used normalisation for data uniformity\u0026mdash;a crucial step for ML task; This process harmonised the ranges of each attribute, ensuring equal contributions to the machine learning algorithms\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e3.3. Feature Selection\u003c/h2\u003e\u003cp\u003eData quality is the most significant contributor for the efficiency of machine learning algorithm\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. We used Random Forest Classifier\u0026mdash;Recursive Feature Elimination using five-fold cross-validation to select the most important features (features that contribute to output prediction) from the data\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e; the process recursively remove the features with the least significance until performance on the validation folds starts decreasing. A vital advantage of this algorithm is that the users don\u0026rsquo;t have to specify desired number of features in advance. The algorithm automatically selects the set of features that results in highest performance during cross-validation.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003e3.4. Building and Training ML Models\u003c/h2\u003e\u003cp\u003eWe compared five machine learning classifiers, namely: Adaboost, Decision Tree Classifier, Logistic Regression, Random Forest Classifier, and Support Vector Classifier using ML-GUIDE. We used these classifiers as the same were used and reported by the developer of ML models for Pima India Diabetes data. Modelling process is described in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The ML models are described in appendix-8 for readers perusal.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e3.5. Hyperparameter Optimization\u003c/h2\u003e\u003cp\u003eHyperparameter optimisation is one of the most critical tasks in building efficient ML models\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. We used GridSearchCV with 5-fold cross-validation to find the best combination of hyperparameters; currently it is the only option available in the software for hyperparameter optimisation. Using the ML-GUIDE software, we selected the best models from GridSearchCV for final training.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e3.6. Testing and Evaluation\u003c/h2\u003e\u003cp\u003eWe used \u0026ldquo;train_test_split\u0026rdquo; (train:test ratio 75:25) function in the software to partition Pima data into training and testing data. The test dataset, comprising a total of 192 participants (including diabetic and non-diabetic) were used to assess the efficacy of the trained models. The diabetic:non-diabetic participant ratio is same as original dataset. The models were evaluated using Accuracy, Average Precision, and F1 score. These metrics were instrumental in quantitatively gauging the models' predictive prowess; all these metrics are calculated based on True Positives (TP), False Positives (FP), True Negatives (TN), and False Negatives (FN). A\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{c}\\text{c}\\text{u}\\text{r}\\text{a}\\text{c}\\text{y}=\\:\\frac{\\text{T}\\text{P}+\\text{T}\\text{N}}{\\text{T}\\text{P}+\\text{F}\\text{P}+\\text{T}\\text{N}+\\text{F}\\text{N}}\\)\u003c/span\u003e\u003c/span\u003e, F1 score - It is the harmonic mean of the recall (sensitivity) and precision (PPV),\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\:\\text{F}1\\:\\text{s}\\text{c}\\text{o}\\text{r}\\text{e}=2\\text{*}\\frac{\\text{P}\\text{r}\\text{e}\\text{c}\\text{i}\\text{s}\\text{i}\\text{o}\\text{n}\\text{*}\\text{R}\\text{e}\\text{c}\\text{a}\\text{l}\\text{l}}{\\text{P}\\text{r}\\text{e}\\text{c}\\text{i}\\text{s}\\text{i}\\text{o}\\text{n}+\\:\\text{R}\\text{e}\\text{c}\\text{a}\\text{l}\\text{l}}\\)\u003c/span\u003e\u003c/span\u003e. Figure-2 demonstrates the process from data selection to model evaluation used by authors.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003e3.7. Model Deployment\u003c/h2\u003e\u003cp\u003eThe ML-GUIDE software saves the trained models as \u0026ldquo;.pkl\u0026rdquo; files (example: \u0026ldquo;AdaBoostClassifier.pkl\u0026rdquo;). To use trained models, researchers can copy the python script from supplementary file (Appendix-5)\u0026mdash;paste this code into a new empty text file and save the file with the name \u0026ldquo;model.py\u0026rdquo; in the same folder where trained models are saved. After saving the \u0026ldquo;model.py\u0026rdquo; file, follow the steps:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eInstall the Anaconda package.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eAfter installation, open the \"Anaconda Prompt\".\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eTo install the Streamlit app, type \"pip install streamlit\" in \"Anaconda Prompt\".\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eGo to anaconda prompt and navigate to \"model.py\" directory using the \u0026ldquo;cd\u0026rdquo; command. For example, \u0026ldquo;cd models\u0026rdquo; to navigate to the \u0026ldquo;models\u0026rdquo; folder.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eThe command \"streamlit run model.py\" in \"Anaconda Prompt\" will launch the app in default web browser.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eSpecify the model path, for example, \"E:/models/AdaBoostClassifier.pkl\". To get the model path, right click on the model.pkl file and click properties. The path is given in the \u0026ldquo;Location\u0026rdquo; field. Copy the path and add model name following the forward slash (/), for example, \u0026ldquo;/AdaBoostClassifier.pkl\u0026rdquo;\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eFollow the on-screen instructions, ensuring to include all the data preprocessing steps applied during model training.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003cp\u003eFollowing these steps the researchers may use the trained models in their computer for making predictions. The users may also watch the video provided with the manuscript for the same.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003e3.8 Hardware used in this study\u0026ndash;\u003c/strong\u003eWe used a Windows 11 machine with Intel\u003csup\u003e(R)\u003c/sup\u003eCore\u003csup\u003e\u0026trade;\u003c/sup\u003e i5-8250U CPU @ 1.60GHz CPU and 16 GB RAM for testing of software. However, we have also tested the software on Windows 10 machine with Intel\u003csup\u003e(R)\u003c/sup\u003eCore\u003csup\u003e\u0026trade;\u003c/sup\u003e i5-8250U CPU @ 1.60GHz CPU and 8 GB RAM. Currently the software runs on Windows 10 and Windows 11 operating systems only.\u003c/p\u003e"},{"header":"4. Results","content":"\u003cp\u003eWe assessed the predictive performance of five ML algorithms developed and trained using the ML-GUIDE software for binary classification tasks\u0026mdash;diabetes prediction. To evaluate the performance of ML-GUIDE, we reported the hyperparameter optimisation time, training time and model evaluation metrics (accuracy, F1 score and average precision) in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. Notably, in the context of Pima Indians Diabetes dataset, we observed that decision tree and logistic regression algorithms exhibited the shortest times for hyperparameter optimisation and less than 1 second for training of all models. The Random Forest algorithm took the longest time\u0026mdash;242 seconds (almost double\u0026mdash;4-minute) for hyperparameter optimisation when compared to Adaboost (129 seconds\u0026mdash;2 minute). It's important to note that the time taken by ML models is influenced by various factors, including the model's complexity, dataset size, and the processing power of the machine\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\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\u003eFeatures of free open source no-code ML tools compared to the ML-GUIDE software\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"10\"\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\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eS. no.\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTool\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"3\" nameend=\"c5\" namest=\"c3\"\u003e\u003cp\u003eNumber of algorithms provided in the tool\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eInput type\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u003cp\u003eData pre-processing\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c9\"\u003e\u003cp\u003eFeature Selection/ Extraction\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c10\"\u003e\u003cp\u003eData Visualisation\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eClassification\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003eRegression\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cb\u003eML-GUIDE\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e\u003cb\u003e25\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e40\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003e\u003cb\u003e\u0026ldquo;CSV\u0026rdquo;\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u003cb\u003e\u0026radic;\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eENNGene\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003eBED genomic interval files\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHDG-select\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003eGEO \u0026ldquo;SOFT\u0026rdquo; and \u0026ldquo;CSV\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eARX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003e\u0026ldquo;CSV\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMetaboAnalyst\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003eWide variety of formats for metabolomics data\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFeAture Explorer (FAE)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003e\u0026ldquo;CSV\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eARTMO's MLCA toolbox\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e19\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003e\u0026ldquo;txt\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eX\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePDAUG\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003ePeptide data\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMVPANI\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003e\u0026ldquo;txt\u0026rdquo;, \u0026ldquo;xlsx\u0026rdquo;, and \u0026ldquo;mat\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e\u0026radic;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003eX\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\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\u003ePerformance evaluation of ML-GUIDE software for developing ML models on the PIMA Indians Diabetes dataset\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"8\"\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\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eS. no.\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eML Model\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003eHyperparameter optimisation time\u003c/p\u003e\u003cp\u003e(seconds)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTraining time\u003c/p\u003e\u003cp\u003e(seconds)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eAccuracy (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eF1 score\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003eAverage Precision\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eAdaboost\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e129\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e75.52\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.83\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.74\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eDecision Tree\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e71.35\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.79\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.74\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eLogistic Regression\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e78.65\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0.85\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.78\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eRandom Forest\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e242\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e77.60\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.83\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u003cb\u003e0.79\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eSupport Vector Machine\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e77.08\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.84\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.77\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\u003eLogistic Regression emerged as the best performer among the evaluated models, demonstrating top performance across multiple key metrics. Logistic Regression achieved the highest scores in terms of Accuracy and F1 Score. Random Forest Classifier outperformed all the other models on Average Precision metrics. Equally noteworthy are the competitive performances exhibited by the AdaBoost Classifier and the Support Vector Machine. Decision Tree was the worst performer among all the models.\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e represents the comparison of ML models\u0026mdash;automated coding off all the models in the current study with manually developed models on PIMA Indians diabetes dataset \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan additionalcitationids=\"CR33 CR34 CR35 CR36 CR37 CR38 CR39\" citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Please refer appendix-6 for table representing accuracies of different models. The results demonstrate promising performance across multiple models compared to published results. Notably, the AdaBoost Classifier achieved an accuracy of 75.52%, closely approaching the literature-reported accuracy of 78.3%. The DecisionTree Classifier also displayed competitive results, with an accuracy of 71.35%, surpassing some of the reported literature values. Logistic Regression and the Support Vector Machine algorithms outperformed the published results with an accuracy of 78.65% and 77.08% respectively, while the Random Forest Classifier achieved an accuracy of 77.60%, maintaining competitive performance.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"5. Discussion","content":"\u003cp\u003e We developed the ML-GUIDE software for medical researchers to expedite development, training, and testing of machine learning models by eliminating manual algorithm coding process. The surge in medical data prompts researchers to use machine learning for improved patient care \u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e,\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. The coding prerequisites, patient load, and time and resource constraints, however, hinder health experts from experimenting with various ML techniques \u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. To mitigate these challenges, we had three objectives, first, to develop the no-code ML tool for researchers, second, to test the efficacy of this tool, and third, to facilitate researchers in deploying their trained models into practice.\u003c/p\u003e\u003cp\u003eThere are multiple advantages of using ML-GUIDE tool compared to slow, manual, and expert driven algorithmic process. The tool offers a user-friendly Graphical User Interface (GUI), enabling researchers to experiment with machine learning (ML) algorithms without learning to code. Unlike the limited diversity of ML models in open-source no-code tools available on PubMed \u003csup\u003e\u003cspan additionalcitationids=\"CR10 CR11 CR12 CR13 CR14 CR15\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e,\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e, the ML-GUIDE software provides a comprehensive selection of ML algorithms, encompassing 25 classification and 40 regression algorithms. Additionally, it allows users to manually adjust key hyperparameters or utilize hyperparameter optimization techniques to fine-tune model performance. Ensuring the reliability and generalizability of trained ML models requires testing on external datasets. The ML-GUIDE software simplifies this process with a one-click option for testing trained models on external datasets. Notably, the increasing use of ML techniques in healthcare mirrors the earlier adoption of statistical software like SPSS and STATA. Just as these platforms made advanced statistical methods accessible to non-statisticians, ML-GUIDE aims to provide access to state-of-the-art ML algorithms to non-coders in clinical and research settings.\u003c/p\u003e\u003cp\u003eTo assess the software's effectiveness, we developed five ML models using ML-GUIDE and compared their accuracy with published literature \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan additionalcitationids=\"CR33 CR34 CR35 CR36 CR37 CR38 CR39\" citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. We used Pima Indians Diabetes dataset to train these models. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e illustrates that the predictive accuracy of ML models generated with ML-GUIDE aligns with those published in the literature. There are minor variations in the accuracy of diabetes prediction among models from different studies. The variability (minimum to maximum) in accuracy of models in literature compared to models developed in this study ranges from 0.15% to 7.54%; this can be attributed to variations in data preprocessing, feature selection, and hyperparameter optimisation techniques used in different studies \u003csup\u003e\u003cspan additionalcitationids=\"CR46\" citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. Our results confirm that the ML models created using the ML-GUIDE software perform at par with the models found in existing literature. Moreover, tests on the Pima Indians diabetes dataset also showed that each model (depending upon processor speed) can be trained in less than one second using 576 samples. The limited sample size does not strongly emphasize the speed of model training in the software, but it does indicate that ML-GUIDE operates within a reasonable timeframe.\u003c/p\u003e\u003cp\u003eAny ML model regardless of their training is only useful when deployed into practice; the crucial test is whether they improve patient outcomes or not \u003csup\u003e\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e. We couldn't find any software articles in our review \u003csup\u003e\u003cspan additionalcitationids=\"CR10 CR11 CR12 CR13 CR14 CR15\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e,\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e that helped researchers deploy their machine learning models. To deploy and use the trained models the ML-GUIDE software enables users to save the trained models as \u0026ldquo;\u0026rsquo;pkl\u0026rdquo; files and also save a \"txt\" file documenting the model development steps (pipeline) for future reference. We provided a Python script as guidance for researchers to deploy their trained ML models to test them in real-time clinical settings. This code can also be utilised for deploying models on an online server to build web applications. However, we have not provided the procedure to integrate trained machine learning models into web applications as it is beyond the scope of the current article.\u003c/p\u003e\u003cp\u003eOne of the noteworthy aspects of the ML-GUIDE tool is its ability to make machine learning more accessible, especially for non-programmers. This inclusiveness allows a broader community of healthcare researchers to engage in machine learning exploration for addressing complex healthcare challenges. We encourage the researchers to extensively explore the software's capabilities and test its limitations.\u003c/p\u003e"},{"header":"6. Conclusion","content":"\u003cp\u003eMany researchers intend to use machine learning in healthcare, but lack of time, programming expertise, and resources slows them down. To facilitate healthcare researchers, we created ML-GUIDE\u0026mdash;a simple tool that lets users quickly build ML models with a point-and-click interface. By eliminating the need for programming: a significant barrier for non-programmers, ML-GUIDE empowers researchers to engage in development of ML-based applications. This open-access tool offers a wide range of algorithms, data preprocessing methods, and hyperparameter optimization options. Our evaluation shows that ML-GUIDE produces models with predictive accuracy on par with those found in existing literature.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics declaration:\u0026nbsp;\u003c/strong\u003eNot applicable\u003cstrong\u003e.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors declare no financial or non-financial competing interests.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Sharing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eThe study is conducted on the publicly available Pima Indians Diabetes Dataset, which can be downloaded from the Kaggle website:\u0026nbsp;\u003c/em\u003ehttps://www.kaggle.com/datasets/uciml/pima-indians-diabetes-database\u003cem\u003e.\u003c/em\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBigML. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://bigml.com/\u003c/span\u003e\u003cspan address=\"https://bigml.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eD Mukunthu \u0026amp; S Gillett. Announcing automated ML capability in azure machine learning. (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://azure.microsoft.com/en-us/blog/announcing-automated-ml-capability-in-azure-machine-learning/\u003c/span\u003e\u003cspan address=\"https://azure.microsoft.com/en-us/blog/announcing-automated-ml-capability-in-azure-machine-learning/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAmazon SageMaker. (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://aws.amazon.com/sagemaker/\u003c/span\u003e\u003cspan address=\"https://aws.amazon.com/sagemaker/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCloud, A. M. L. \u003cem\u003eCloud AutoML\u003c/em\u003e (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cloud.google.com/automl/\u003c/span\u003e\u003cspan address=\"https://cloud.google.com/automl/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRapidMiner \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://rapidminer.com/\u003c/span\u003e\u003cspan address=\"https://rapidminer.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDataRobot Automated Machine Learning. (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.datarobot.com/platform/automated-machine-learning/\u003c/span\u003e\u003cspan address=\"https://www.datarobot.com/platform/automated-machine-learning/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang, Z., Wang, Y., Wang, C., Chen, J. \u0026amp; Zheng, Z. LLM Hallucinations in Practical Code Generation: Phenomena, Mechanism, and Mitigation. Preprint at (2025). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.48550/arXiv.2409.20550\u003c/span\u003e\u003cspan address=\"10.48550/arXiv.2409.20550\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlami, H. et al. Understanding the integration of artificial intelligence in healthcare organisations and systems through the NASSS framework: a qualitative study in a leading Canadian academic centre. \u003cem\u003eBMC Health Serv. Res.\u003c/em\u003e \u003cb\u003e24\u003c/b\u003e, 701 (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChalupov\u0026aacute;, E. et al. ENNGene: an Easy Neural Network model building tool for Genomics. \u003cem\u003eBMC Genom.\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e, 248 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHameed, S. S., Hassan, R., Hassan, W. H., Muhammadsharif, F. F. \u0026amp; Latiff, L. A. HDG-select: A novel GUI based application for gene selection and classification in high dimensional datasets. \u003cem\u003ePloS One\u003c/em\u003e. \u003cb\u003e16\u003c/b\u003e, e0246039 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEicher, J., Bild, R., Spengler, H., Kuhn, K. A. \u0026amp; Prasser, F. A comprehensive tool for creating and evaluating privacy-preserving biomedical prediction models. \u003cem\u003eBMC Med. Inf. Decis. Mak.\u003c/em\u003e \u003cb\u003e20\u003c/b\u003e, 29 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXia, J., Psychogios, N., Young, N. \u0026amp; Wishart, D. S. MetaboAnalyst: a web server for metabolomic data analysis and interpretation. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cb\u003e37\u003c/b\u003e, W652\u0026ndash;660 (2009).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSong, Y. et al. FeAture Explorer (FAE): A tool for developing and comparing radiomics models. \u003cem\u003ePloS One\u003c/em\u003e. \u003cb\u003e15\u003c/b\u003e, e0237587 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAghababaei, M. et al. Introducing ARTMO\u0026rsquo;s Machine-Learning Classification Algorithms Toolbox: Application to Plant-Type Detection in a Semi-Steppe Iranian Landscape. \u003cem\u003eRemote Sens.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e, 4452 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJoshi, J. \u0026amp; Blankenberg, D. PDAUG: a Galaxy based toolset for peptide library analysis, visualization, and machine learning modeling. \u003cem\u003eBMC Bioinform.\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e, 197 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePeng, Y. et al. MVPANI: A Toolkit With Friendly Graphical User Interface for Multivariate Pattern Analysis of Neuroimaging Data. \u003cem\u003eFront. Neurosci.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e, 545 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKamal, K., Sharma, M., Rahul, M., Rakesh, K. \u0026amp; Koushal, V. Ensconcing a biostatistics clinic in tertiary care research institute of India: A descriptive study. \u003cem\u003eJ. Fam Med. Prim. Care\u003c/em\u003e. \u003cb\u003e11\u003c/b\u003e, 2051 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKorot, E. et al. Code-free deep learning for multi-modality medical image classification. \u003cem\u003eNat. Mach. Intell.\u003c/em\u003e \u003cb\u003e3\u003c/b\u003e, 288\u0026ndash;298 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWaring, J., Lindvall, C. \u0026amp; Umeton, R. Automated machine learning: Review of the state-of-the-art and opportunities for healthcare. \u003cem\u003eArtif. Intell. Med.\u003c/em\u003e \u003cb\u003e104\u003c/b\u003e, 101822 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIBM. IBM SPSS.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSTATA.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEric Siegel. Models Are Rarely Deployed: An Industry-wide Failure in Machine Learning Leadership. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.kdnuggets.com/2022/01/models-rarely-deployed-industrywide-failure-machine-learning-leadership.html\u003c/span\u003e\u003cspan address=\"https://www.kdnuggets.com/2022/01/models-rarely-deployed-industrywide-failure-machine-learning-leadership.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePima Indians Diabetes Database - Kaggle.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDomingos, P. A few useful things to know about machine learning. \u003cem\u003eCommun. ACM\u003c/em\u003e. \u003cb\u003e55\u003c/b\u003e, 78\u0026ndash;87 (2012).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePedregosa, F. et al. Scikit-learn: Machine Learning in Python. Preprint at (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://arxiv.org/abs/1201.0490\u003c/span\u003e\u003cspan address=\"http://arxiv.org/abs/1201.0490\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePima Indians Diabetes. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://data.world/uci/pima-indians-diabetes\u003c/span\u003e\u003cspan address=\"https://data.world/uci/pima-indians-diabetes\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKumari, S., Kumar, D. \u0026amp; Mittal, M. An ensemble approach for classification and prediction of diabetes mellitus using soft voting classifier. \u003cem\u003eInt. J. Cogn. Comput. Eng.\u003c/em\u003e \u003cb\u003e2\u003c/b\u003e, 40\u0026ndash;46 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi, W. \u0026amp; Liu, Z. A method of SVM with Normalization in Intrusion Detection. \u003cem\u003eProcedia Environ. Sci.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, 256\u0026ndash;262 (2011).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDarst, B. F., Malecki, K. C. \u0026amp; Engelman, C. D. Using recursive feature elimination in random forest to account for correlated variables in high dimensional data. \u003cem\u003eBMC Genet.\u003c/em\u003e \u003cb\u003e19\u003c/b\u003e, 65 (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFrank \u0026amp; Hutter Holger Hoos, \u0026amp; Kevin Leyton-Brown. An efficient approach for assessing hyperparameter importance. (2014).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eL\u0026rsquo;Heureux, A., Grolinger, K., Elyamany, H. F. \u0026amp; Capretz, M. A. M. Machine Learning With Big Data: Challenges and Approaches. \u003cem\u003eIEEE Access.\u003c/em\u003e \u003cb\u003e5\u003c/b\u003e, 7776\u0026ndash;7797 (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDhilsath Fathima, M. \u0026amp; Justin Samuel, S. Improved Adaboost Algorithm with Regression Imputation for Prediction of Chronic Type 2 Diabetes Mellitus. in Communication and Intelligent Systems (eds Sharma, H., Gupta, M. K., Tomar, G. S. \u0026amp; Lipo, W.) (, vol 204 691\u0026ndash;708 (Springer Singapore, Singapore, (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBozkurt, M. R., Yurtay, N., Yilmaz, Z. \u0026amp; Sertkaya, C. Comparison of different methods for determining diabetes. \u003cem\u003eTurk. J. Electr. Eng. Comput. Sci.\u003c/em\u003e \u003cb\u003e22\u003c/b\u003e, 1044\u0026ndash;1055 (2014).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChatrati, S. P. et al. Smart home health monitoring system for predicting type 2 diabetes and hypertension. \u003cem\u003eJ. King Saud Univ. - Comput. Inf. Sci.\u003c/em\u003e \u003cb\u003e34\u003c/b\u003e, 862\u0026ndash;870 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDeng, D. \u0026amp; Kasabov, N. On-line pattern analysis by evolving self-organizing maps. \u003cem\u003eNeurocomputing\u003c/em\u003e \u003cb\u003e51\u003c/b\u003e, 87\u0026ndash;103 (2003).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChang, V., Bailey, J., Xu, Q. A. \u0026amp; Sun, Z. Pima Indians diabetes mellitus classification based on machine learning (ML) algorithms. \u003cem\u003eNeural Comput. Appl.\u003c/em\u003e \u003cb\u003e35\u003c/b\u003e, 16157\u0026ndash;16173 (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRajendra, P. \u0026amp; Latifi, S. Prediction of diabetes using logistic regression and ensemble techniques. \u003cem\u003eComput. Methods Programs Biomed. Update\u003c/em\u003e. \u003cb\u003e1\u003c/b\u003e, 100032 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSaxena, R., Sharma, S. K., Gupta, M. \u0026amp; Sampada, G. C. A Novel Approach for Feature Selection and Classification of Diabetes Mellitus: Machine Learning Methods. \u003cem\u003eComput. Intell. Neurosci.\u003c/em\u003e 1\u0026ndash;11 (2022). (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZou, Q. et al. Predicting Diabetes Mellitus With Machine Learning Techniques. \u003cem\u003eFront. Genet.\u003c/em\u003e \u003cb\u003e9\u003c/b\u003e, 515 (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKarthikeyani, V., Parvin Begum, I. \u0026amp; Tajudin, K. Shahina Begam, I. Comparative of Data Mining Classification Algorithm (CDMCA) in Diabetes Disease Prediction. \u003cem\u003eInt. J. Comput. Appl.\u003c/em\u003e \u003cb\u003e60\u003c/b\u003e, 26\u0026ndash;31 (2012).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBaro, E., Degoul, S., Beuscart, R. \u0026amp; Chazard, E. Toward a Literature-Driven Definition of Big Data in Healthcare. \u003cem\u003eBioMed Res. Int.\u003c/em\u003e 1\u0026ndash;9 (2015). (2015).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHay, S. I., George, D. B., Moyes, C. L. \u0026amp; Brownstein, J. S. Big data opportunities for global infectious disease surveillance. \u003cem\u003ePLoS Med.\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e, e1001413 (2013).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRazeghi, O. et al. CemrgApp: An interactive medical imaging application with image processing, computer vision, and machine learning toolkits for cardiovascular research. \u003cem\u003eSoftwareX\u003c/em\u003e \u003cb\u003e12\u003c/b\u003e, 100570 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMattonen, S. A. et al. Quantitative imaging feature pipeline: a web-based tool for utilizing, sharing, and building image-processing pipelines. \u003cem\u003eJ. Med. Imaging Bellingham Wash.\u003c/em\u003e \u003cb\u003e7\u003c/b\u003e, 042803 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAhsan, M., Mahmud, M., Saha, P., Gupta, K. \u0026amp; Siddique, Z. Effect of Data Scaling Methods on Machine Learning Algorithms and Model Performance. \u003cem\u003eTechnologies\u003c/em\u003e \u003cb\u003e9\u003c/b\u003e, 52 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEjiro Onose. How to Solve Reproducibility in ML. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://neptune.ai/blog/how-to-solve-reproducibility-in-ml\u003c/span\u003e\u003cspan address=\"https://neptune.ai/blog/how-to-solve-reproducibility-in-ml\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGhorbani, R. \u0026amp; Ghousi, R. Comparing Different Resampling Methods in Predicting Students\u0026rsquo; Performance Using Machine Learning Techniques. \u003cem\u003eIEEE Access.\u003c/em\u003e \u003cb\u003e8\u003c/b\u003e, 67899\u0026ndash;67911 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKelly, C. J., Karthikesalingam, A., Suleyman, M., Corrado, G. \u0026amp; King, D. Key challenges for delivering clinical impact with artificial intelligence. \u003cem\u003eBMC Med.\u003c/em\u003e \u003cb\u003e17\u003c/b\u003e, 195 (2019).\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":"","lastPublishedDoi":"10.21203/rs.3.rs-8255495/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8255495/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMachine learning (ML) is transforming healthcare research, but developing models can be complex and time-consuming. We created ML-GUIDE, a no-code tool with a graphical interface, to make ML model building faster and more accessible.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe developed the ML-GUIDE software using Python language and the Model-View-Controller architecture. ML-GUIDE provides 40 regression and 25 classification algorithms. We used publicly available Pima Indians Diabetes dataset to assess efficacy of the software. As per literature on ML models for Pima dataset, we developed, trained, and compared results from five diabetes prediction models: Adaboost, Decision Tree, Logistic Regression, Random Forest Classifier, and Support Vector Classifier for this study. We used metrics such as accuracy, average precision, and F1 score to evaluate the performances of ML models.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe found that the model developed using ML-GUIDE software performed similar to published models. The variability (minimum to maximum) in accuracy of models in literature compared to models developed in this study ranges from 0.15% to 7.54%. The study also revealed that logistic regression with accuracy (78.65%), F1 score (0.849), and average precision (0.776) outperformed other models in predicting diabetes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe results demonstrated that ML-GUIDE—a coding free tool matches performance of manual, time-consuming and expert driven algorithms. Therefore, ML-GUIDE tool can be used to rapidly test, develop and deploy ML-based healthcare solutions.\u003c/p\u003e","manuscriptTitle":"Training and Testing Healthcare Models using Machine Learning Graphical User Interface Development Environment (ML-GUIDE) Software","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-12 14:29:33","doi":"10.21203/rs.3.rs-8255495/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":"aa53b02f-f412-4d43-ab87-19497937003c","owner":[],"postedDate":"December 12th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":59498999,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":59499000,"name":"Health sciences/Health care"},{"id":59499001,"name":"Physical sciences/Mathematics and computing"},{"id":59499002,"name":"Health sciences/Medical research"}],"tags":[],"updatedAt":"2025-12-16T08:38:33+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-12 14:29:33","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8255495","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8255495","identity":"rs-8255495","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}