International Journal of Computer Techniques Volume 12 Issue 4 | A Comparative Study of Predictive Modelling of Diabetes Using Machine Learning Algorithms
A Comparative Study of Predictive Modelling of Diabetes Using Machine Learning Algorithms
Authors:
Vishal Kumar, Varun Bansal, Naman Saini, Arun Saini
Department of Computer Science Engineering, Shobhit University, Gangoh, India
Emails: Vishal.kumar@shobhituniversity.ac.in, Varun.bansal@shobhituniversity.ac.in, Naman.saini@shobhituniversity.ac.in, Arun.saini@shobhituniversity.ac.in
Journal: International Journal of Computer Techniques (IJCT)
Volume: 12 | Issue: 4 | Publication Date: July – August 2025
ISSN: 2394-2231 | Journal URL: https://ijctjournal.org/
Abstract
This paper presents a comparative study of machine learning algorithms for diabetes prediction. Using the PIMA Indian Diabetes dataset, ensemble models combining Logistic Regression, Decision Tree, and ANN achieved high accuracy. Advanced techniques like GA-XGBoost and SMOTEENN addressed class imbalance. SHAP-based interpretation improved model transparency. The study also explored cardiovascular prediction and heart sound classification, highlighting the importance of robust algorithms and real-world data integration.
Keywords
Diabetes, Type 1 Diabetes, Type 2 Diabetes, Gestational Diabetes, Insulin, Glucose, Machine Learning, XGBoost, SHAP, SMOTEENN, Ensemble Models
Conclusion
Ensemble and deep learning models significantly improve diabetes prediction accuracy. SHAP interpretation enhances clinical relevance. Future work should focus on multi-disease classification and real-time medical data integration to improve generalizability and diagnostic reliability.
References
Includes 15 references from PLoS ONE, MDPI, arXiv, CDC, Nature, and biomedical engineering journals covering diabetes prediction, ensemble learning, and model interpretability.
