Heart Disease Prediction using SVM and RF

Heart disease is one of the top priority health issues in the world, and early detection is paramount to enhancing the efficacy of treatment and reducing complications. Recently, machine learning has shown promises in predicting heart disease. The present study examines predictions based on random forests, and support vector machines on realistic data sets. The Boruta algorithm is used to find all the significant attributes from the dataset with respect to an outcome variable. From 13 input attributes, 6 attributes are selected including age, chest pain, maximum heart rate, ST depression induced by exercise relative to rest, number of major vessels colored by flourosopy, Thallium stress test. In this research paper, the presence of heart disease is predicted by employing Random Forest, Support Vector Machine. The F1- Score, recall and precision score of Random forest classifiers are 84.60%, 85.12% and 84.82%, respectively.