AN EFFICIENT HEART DISEASE PREDICTION USING SELECT K-BEST FEATURES AND MACHINE LEARNING ALGORITHMS

Cardiovascular diseases (CVDs) remain a leading cause of global mortality, emphasizing the urgent need foraccurate and early diagnosis. This project presents an efficient heart disease prediction model that leveragesmachine learning techniques combined with K-Best feature selection to enhance predictive accuracy. A dataset ofclinical patient records is preprocessed and analysed using six supervised learning algorithms: Decision Tree, KNearest Neighbours, Logistic Regression, Support Vector Machine, Random Forest, and XGBoost. The XGBoostclassifier, coupled with k-fold cross-validation, demonstrated superior performance with an accuracy exceeding96%. A user-friendly GUI built with Tainter enables real-time prediction based on medical inputs, bridging thegap between machine learning models and practical healthcare applications. The results confirm the viability ofthe proposed hybrid model as a clinical decision-support tool for early and accurate heart disease detection.