HEART DISEASE PREDICTION USING MACHINE LEARNING

Authors:

Mohammad Mustufa Anis Vindhani [1], Amit Bhusari [2]

Department of MCA, Trinity Academy of Engineering, Pune, India

Assistant Professor Department of MCA,Trinity Academy of Engineering, Pune, India

Abstract: This project presents a robust machine learning system designed to predict heart disease using the UCI Heart Disease dataset, which comprises303 patient records and 14clinical features, including age, chest pain type, and maximum heart rate. Through comprehensive exploratory data anal-lysis (EDA), significant correlations were identified, particularly with chest pain type (correlation: 0.43) and maximum heartrate(correlation: 0.42),guiding feature prioritization. Three machine learning models Logistic Regression, K-Nearest Neighbours (KNN), and Random Forest—were de-eloped, trained, and rigorously evaluated. Logistic Regression achieved the highest test accuracy of 88.52%, with cross-validated metrics demon-strating reliability: precision (82.08%), recall (92.12%), F1-score (86.73%),and ROC-AUC ( 0.95).