Hybrid Model Using Feature Selection and Classifier in Big data Healthcare Analytics

Abstract

AbstractThe growing era of technology in healthcare results in a large amount of data generation termed big data. Mining this data generating valuable insights is crucial in the classification or prediction of disease at an early stage. Effective decision making in the domain of healthcare happens using advanced data mining techniques. In this article, a hybrid model using feature selection and classifier is examined. The feature selection method is applied to the heart disease dataset to select more appropriate features and results given to the classifier to predict the heart disease at an early stage. Three feature selection methods analysis of variance, Pearson’s correlation coefficient, and mutual information gain applied on heart disease dataset and the performance of random forest classifier were examined over heart disease prediction. Experimental results showed that the feature selection approach increases the accuracy of the classifier.KeywordsBig dataData miningFeature selectionHealthcareMachine learning