Predictive Analytics for Heart Disease Detection: A Machine Learning Approach
S Praveena Rachel Kamala, S. Gayathri, N.Muthuvairavan Pillai, L. A. Anto Gracious, C M Varun, R. Sıva Subramanıan
Abstract
One of the most life-threatening diseases that causes human beings to suffer is heart disease. According to WHO report, more than 17.9 million people are affected and raised to death due to cardiovascular diseases around the globe. Many factors that influence the disease are high blood pressure, LDL cholesterol, obesity, physical inactivity, diabetes, etc. An earlier diagnosis of the disease helps to change the lifestyle and the start of treatment helps to save the human being life. Machine learning techniques are used in this work to analyze heart disease effectively. The purpose of the research is to perform efficient heart disease analysis using ML models. Machine learning is a sub-category of AI that makes decisions based on past and historical data. ML can be sub-classified into 3 types. In this research, supervised ML models are considered for analysis. In supervised ML models: K-NN, SVM, and RF are applied. The experimental approach is conducted with the heart disease dataset with three different ML models. The results demonstrate that the RF approach has the best performance across all metrics, subsequent to the SVM model and the KNN model. Among the three models, the RF model has the greatest sensitivity (0.8654), precision (0.8182), specificity (0.7674), and accuracy (0.82105) compared to other approaches in this particular heart disease dataset.