Explainable Artificial Intelligence for Early Prediction of Cardiovascular Diseases Using Multi-Model Clinical data

Abstract

Cardiovascular diseases (CVDs) continue to be one of the biggest public health problems around the world and early detection of risk is crucial for early clinical intervention and prevention. Several machine learning models such as Logistic Regression, Random Forest, Support Vector Machine, XGBoost and a neural network model are comparatively analysed and assessed to predict the performance. In order to enhance the robustness of the model, feature preprocessing, feature selection, class balancing, hyperparameter optimization and cross-validation are included. Explainable AI techniques, such as SHAP and LIME, are used to find global and patient-specific predictors and to give explanations that are interpretable to patients for individual risk predictions.