On October 21, 2026, MSc candidate Sanja Lazić successfully defended her thesis entitled “Interpretable Machine Learning and Deep Learning Models for Type 2 Diabetes Risk Prediction Based on Behavioural and Clinical Indicators” within the Artificial Intelligence Master’s programme at the University of Donja Gorica, Faculty for Information Systems and Technologies, under the mentorship of Prof. Dr. Tomo Popović.
The research was partially supported by the AI-AGE project, with additional support from NCC HPC Montenegro within the EuroCC3 project. The thesis explores the application of explainable artificial intelligence to health risk assessment, focusing on the development of reliable and interpretable predictive models that use behavioural, clinical, and demographic indicators without requiring laboratory tests. The work contributes to research and capacity building in medical AI, particularly in model interpretability, reliability, and the potential application of AI-based tools in preventive healthcare.
ABSTRACT – This thesis presents the development and systematic evaluation of machine learning and deep learning models for assessing the risk of type 2 diabetes using questionnaire-based health indicators. The study analysed a dataset containing 253,680 records and 21 predictive variables, comparing five approaches: Logistic Regression, Random Forest, XGBoost, LightGBM, and a deep neural network. Model performance was evaluated in terms of predictive discrimination, probability calibration, and interpretability. While the deep neural network achieved the highest ROC-AUC score (0.7728), LightGBM demonstrated the most balanced overall performance. Explainable AI techniques, including TreeSHAP, Integrated Gradients, and DeepLIFT SHAP, were employed to identify factors contributing to individual predictions. A prototype web application was also developed to demonstrate how estimated risk and its contributing factors could be presented to users. The findings highlight the potential of interpretable AI models to support non-invasive health risk assessment and preventive screening, while emphasizing the need for further validation before clinical application.