Predicting Diabetic Distress and Emotional Burden in Type-2 Diabetes Using Explainable AI

The main objective of this project is to develop an explainable machine learning system capable of predicting diabetic distress levels in individuals with Type-2 Diabetes. This is achieved by utilizing both clinical and emotional features to classify distress into two categories: low or high. To accomplish this goal, the project first focuses on preprocessing and preparing the dataset to ensure high-quality inputs for effective model training. Following this, a range of existing machine learning algorithms including XGBoost, Random Forest, Support Vector Machine (SVM), Logistic Regression, Ridge Regression, Lasso Regression, and Multi-Layer Perceptron are implemented and compared for their performance in predicting distress levels. Additionally, the project explores the development of advanced ensemble models such as Stacking Classifier, Voting Classifier, TabNet, and CatBoost to further enhance prediction accuracy and model interpretability.