Predicting Heart Disease Using Machine Learning: An Evaluation of Logistic Regression, Random Forest, SVM, and KNN Models on the UCI Heart Disease Dataset

This study evaluates the performance of three machine learning models—Random Forest, Support Vector Machine (SVM), and Logistic Regression—in predicting heart disease using the "Heart Disease UCI" dataset from Kaggle. The models were assessed based on accuracy, precision, recall, and F1-score, both with and without feature selection techniques such as Chi-Square and Mutual Information.Without feature selection, Random Forest achieved the highest performance with an accuracy of 89.7%, followed by SVM with 87.0%, and Logistic Regression with 84.2%. Using Mutual Information for feature selection, Random Forest achieved an accuracy of 85.3%, SVM 87.0%, and Logistic Regression 82.6%. With Chi-Square feature selection, Random Forest and Logistic Regression both showed an accuracy of 83.2%, while SVM achieved 82.6%.The results indicate that Random Forest consistently performs well across different scenarios, making it a robust choice for heart disease prediction. Feature selection did not significantly enhance model performance, suggesting that the initial features in the dataset are already highly relevant. These findings highlight the potential of machine learning, especially Random Forest, in aiding clinical diagnosis of heart disease. Further research is needed to validate these models on larger, more diverse datasets and to explore advanced feature selection techniques for improved model performance.