Abstract

The application of machine learning algorithms in medical disease diagnosis and treatment has become increasingly prevalent, particularly in the context of heart disease prediction. With a rising number of sudden heart-related fatalities, accurate prediction and diagnosis have become paramount. Data mining techniques and machine learning algorithms offer significant contributions in this domain, supporting the development of software that assists healthcare professionals in making informed decisions about heart disease risk and diagnosis. In this study, we focus on leveraging data mining techniques to predict heart disease in advance for timely intervention. We employ various algorithms for comparative analysis, and our findings indicate that the Random Forest algorithm, complemented by Support Vector Machines (SVM), yields the highest accuracy in prediction. Using a dataset comprising 303 samples and 14 input features, along with one output feature, we achieved promising results. This dataset, sourced from the UCI Machine Learning Repository, is divided into 65% for training and 35% for testing, yielding a precision of 0.763 and a recall of 0.935 in predicting negative class tuples.

Description