Exploratory Analysis to Predict Heart Disease Occurrence through Machine Learning Approaches
Abstract
Heart disease is one of the leading diseases in the world that affects human life. Prediction of heart disease at early stage with symptoms is crucial to save life. In real fie, diagnosis of heart disease through traditional medical history hasn’t been reliable. Classifying healthy people and people with heart disease can be done with non-invasive-based methods such as machine learning. In the proposed study, machine Learning methods like Decision Tree, Naïve Baye’s, k – Nearest Neighbour, Support Vector Machine (RBF and Linear Kernel), Logistic Regression and Artificial Neural Network are used for the analysis of prediction of Heart Disease patients from normal persons using heart disease dataset from UCI repository. Popular machine learning algorithms are implemented and tested with k – fold cross-validation, and the performances of those methods are evaluated with classifiers performance evaluation metrics such as classification accuracy, specificity, sensitivity. Proposed system can easily identify and classify people with heart disease from healthy people. Naïve Baye’s Shows good performance in overall in terms of Accuracy, Sensitivity and Specificity. Support Vector Machine with RBF kernel and Logistic Regression also show good performance in terms of three performance measure next to Naïve Baye’s method. Decision Tree method gives average performance. The primary motive of this work is to predict heart diseases with high accuracy rate.