Machine Learning and Deep Learning Models for Early Detection of Heart Disease
Gurpreet Singh, Kalpna Guleria, Shagun Sharma
Abstract
The prevalence of cardiovascular disease is a significant worldwide health concern. health concern and is responsible for an increased fatality rate. Effective treatment and prevention of heart disease depend on early detection and precise prediction. This work proposes various machine learning (ML) and deep learning (DL) models for the prediction of heart disease based on a large dataset of patient traits and medical indicators defining the presence of heart disease. This work uses various classifiers namely Naive Bayes, Multilayer Perceptron (MLP), decision tree, and Logistic Regression for developing an effective heart disease prediction model. These models have been fed with a dataset containing 1025 tuples and 11 attributes such as sex, age, thalach, restecg etc. The proposed model has been developed with a train and test ratio of 80:20, The aforementioned phenomenon has led to the emergence of certain measures, including precision, recall, accuracy, and Fl-measure, as indicators of the achieved results. The results have depicted that out of all the ML and DL models, the Decision Tree exhibited the highest accuracy with a rate of 98.04%. However, MLP has been also identified as the second outperforming model by showing 95.51 % for binary heart disease classification. Furthermore, the performance of Naive Bayes and logistic regression have shown an accuracy of 83.12%, and 84.48%, respectively.