Predictive Modelling for Heart Disease Diagnosis: A Comparative Study of Classifiers

Abstract

INTRODUCTION: Cardiovascular diseases, including heart disease, remain a significant cause of morbidity and mortality worldwide. Timely and accurate diagnosis of heart disease is crucial for effective intervention and patient care. With the emergence of machine learning techniques, there is a growing interest in leveraging these methods to enhance diagnostic accuracy and predict disease outcomes. OBJECTIVES: This study evaluates the performance of three machine learning classifiers—Naive Bayes, Logistic Regression, and k-Nearest Neighbors in predicting heart disease based on patient attributes. METHODS: In this study, we explore the application of three prominent machine learning classifiers—Naive Bayes, Logistic Regression, and k-Nearest Neighbors (kNN)—to predict the presence of heart disease based on a set of patient attributes. RESULTS: Using a dataset of 303 patient records with 14 attributes, including age, sex, and cholesterol levels, the data is pre-processed, scaled, and split into training and test sets. Each classifier is trained on the training set and evaluated on the test set. Results reveal that Naive Bayes and k-Nearest Neighbors classifiers outperform Logistic Regression in terms of accuracy, precision, recall, and area under the ROC curve (AUC). CONCLUSION: This study underscores the promising role of machine learning in medical diagnosis, showcasing the potential of Naive Bayes and k-Nearest Neighbors classifiers in improving heart disease prediction accuracy. Future work could explore advanced classifiers and feature selection techniques to enhance predictive accuracy and generalize findings to larger datasets.