An Identification of Heart Disorder Using Machine Learning Algorithms

Abstract

Cardiovascular disease, sheathes numerous conditions that affect the heart which is the predominant reason for death worldwide over the span of the past few decades. Recently, there is an increase in mortality rate of young people ranging from age of 30 to 40. Therefore, this created a perilous need of prediction of cardiovascular disease. This paperpredicts the heart diseasebyexploitation of coronary heart sound and uses machine learning methods such as K-Nearest Neighbour, Gaussian Naive Bayes, Random Forest, Support Vector Machine, Decision Tree, and Logistic Regression. The proposed method uses dataset collected both stable persons and from patients with heart disease were collected around 650 cardiac sound patterns were collected. The proposed work consists of pre-processing, feature extraction and classification. The audio signal of heart is given as the input, then onset detector and onset strengthare used to forpre-processing to remove the unwanted noise. After pre-processing the Mel Frequency Ceptral Coefficients are used to extract the features. The extracted features are given as the input to Decision Tree, K- Nearest Neighbour, Logistic Regression, Naive Bayes Classifier, Support Vector Machine and Random Forest for classification which classifies into five categories namely, normal, artefact, extrahls, extrasystole and murmur. From the experimental results it is found that Random Forest gives the highest accuracy 87. 69 % when comparing with other standard methods.