Early Heart Disease Prediction Using Supervised Machine Learning: A Performance Evaluation on UCI Data

Abstract

Cardiovascular diseases (CVDs) remain a significant public health burden globally and coronary artery disease (CAD) is a major cause of heart attacks. In this paper, we explore the use of machine learning (ML) methods as tools for improving the early detection of CVD. Fifteen supervised ML and deep learning algorithms were used for the study in python 3.10. Model training and evaluation were performed by Scikit-learn, Pandas was used for descriptive statistical analysis and plots were generated with Matplotlib and Seaborn. The dataset utilized, which includes 920 instances, is a merged information from four different locations (Cleveland, Hungary, Switzerland and VA Long Beach) obtained from the UCI Machine Learning Repository. The Histogram-based Gradient Boosting Classifier performed best among the models tested, with an accuracy of 93.5 % and a 5-fold crossvalidation accuracy of 92.4 %. It also obtained high scores in precision, recall, and F1 measurements for both classes and was able to classify people with and without the heart disease. These results highlight the promise of ML for early CVD detection and timely clinical interventions.