Cardiac Abnormality Prediction Using Multiple Machine Learning Approaches

Abstract

Heart disease is one of the deadliest diseases in the modern world. Consistently right around 26 million patients are being influenced by this sort of problem. From the heart consultant and specialist’s perspective, it is intricate to foresee the cardiovascular breakdown correctly. This early detection, in addition, can help control the side effects of the illness just as the appropriate treatment of the Abnormality. Machine learning can play a vital role to predict heart disease using previous medical data. It’s possible to predict heart disease using various data mining and machine learning algorithms, a faster, easier and cost-effective solution for medical science. The primary purpose of this paper is to predict Cardiac Abnormalities using various data mining techniques. 6 machine learning algorithms were used (Decision Tree, K-Nearest Neighbors, Logistic Regression, Naïve Bayes Classifier, Random Forest, Support Vector Machine) to predict whether a person has cardiovascular disease or not. We applied a raw dataset with 12 attributes and our engineered dataset with 16 attributes on these algorithms. In both dataset number of data points were 62500. After analyzing the accuracy, Logistic Regression provides better accuracy, and Decision Tree provides the lowest accuracy.