Prediction of Heart Disease with Different Attributes Combination by Data Mining Algorithms

Abstract

Heart disease is considered the most dangerous and fatal infection in the human body. This globally fatal disease cannot be identified easily by a general practitioner, and it requires an analyst or expert to detect it. In the field of medical science, machine learning plays important role in disease prediction to identify the heart infection features. In this perspective, this research work proposes a new technique to predict heart disease by using various classifier algorithms such as random forest, gradient boosting, support vector machine, and K- nearest neighbor algorithms. For this purpose, the classification accuracy and the obtained results of each predictor have been compared. In each analysis, machine learning classifier algorithms: random forest, gradient boosting, support vector machine, and K-nearest neighbor algorithms are used and finally defect, gradient boosting, which has calculated high accuracy with low error values and high correlation value when compared to other used algorithms.KeywordsFeature selection methods: extra treeRandom forestGradient boostingSupport vector machineK-nearest neighbor algorithmsContingency coefficientAdjusted contingency coefficientCorrelation coefficient and phi coefficientRoot mean square error