Feature Selection for Predicting Heart Disease Using Black Hole Optimization Algorithm and XGBoost Classifier

Cardiovascular disease is the most fatal disease in today modern human certainty. This disease affects a person so rapidly that there is barely any time they need the treatment. Therefore, it is most difficult job for the medical diagnosis to identify patients accurately on a timely basis. For reducing the cardiovascular disease, an effective feature selection technique and classification based prediction technique is adopted. On high-dimensional medical data, an appropriate feature selection technique as black hole optimization (BHO) algorithm is proposed to select the relevant features for diagnosis. Thereafter, obtained features from the medical dataset are given to the XGBoost classifier for performing classification. Our experimental research shows that integration of BHO and XGBoost technique is used to finds reduced subsets and it also enhances the diagnostic accuracy, when compare with the existing classifiers.