Skip to content
Conference Open access

Early Diabetes Prediction using Explainable Ensemble Learning and Boruta–SHAP

Sep 2026 · 2026 10th International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC) · pp. 1100-1106 · 0 citations · 19 references

Abstract

Diabetes is a long-term metabolic disease that has to be diagnosed as soon as possible to lower the risk of serious health issues. Although machine learning has proven to be a successful method for early diabetes prediction, issues such class imbalance, irrelevant information, and limited model interpretability frequently lower prediction reliability. This paper uses a hybrid Boruta–SHAP feature selection strategy to present an explainable ensemble learning framework for diabetes prediction. Following data cleaning and duplication removal, the 96,146 patient records in the publicly accessible Kaggle diabetes prediction dataset were utilized. To resolve class imbalance, the dataset was preprocessed using one-hot encoding and the Synthetic Minority Over-sampling Technique (SMOTE). The most informative clinical and demographic characteristics were found using hybrid Boruta–SHAP feature selection. Accuracy, precision, recall, F1-score, ROC-AUC, Matthews Correlation Coefficient (MCC), and specificity were used to evaluate five ensemble learning models: Random Forest, XGBoost, LightGBM, CatBoost, and Stacking Ensemble. With an accuracy of 96.97%, a ROC-AUC of 97.65%, and an MCC of 79.76%, CatBoost outperformed the other models that were assessed. SHAP analysis was used to improve the interpretability of the model, while calibration analysis, confidence estimation, clinical risk classification, and ablation experiments were used to evaluate dependability. The experimental findings show that the suggested framework is a promising decision-support tool for early diabetes screening in clinical practice since it offers precise, dependable, and comprehensible diabetes prediction.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.