Predicting heart disease using machine learning algorithms

Abstract

Increasing the accuracy of detecting heart disease is widely studied in the field of machine learning. Such a study is intended to prevent large costs in the field of healthcare and is the reason for the misdiagnosis. As a result, various methods of analyzing disease factors were proposed, aimed at reducing differences in the practice of doctors and reducing medical costs and errors. In this study, 6 classification learning algorithms were used, including machine learning methods such as classification Tree, Close neighborhood method, Naive Bayes, random forest tree, and busting methods. These methods were collected by the University of Cleveland.using csv datacet, they were trained to make an effective and accurate prediction of heart disease. In order to increase the predictive capabilities of algorithms, all methods were trained primarily on non-standardized data. A study was conducted on how much data standardization affects the result using the Standard Scaler method. In the course of the study, this method helped algorithms such as KNN and SVC improve the result by about 25\%.

Keywords

• classification
• standardization
• training selection
• metrics
• busting
• confusion matrix