Novel and Efficient Hybrid Model for Classification of Heart Disease

Abstract

To propose an efficient heart disease classification algorithm to predict disease in early stage so that rate of death can be reduced. A hybrid intelligent model of Genetic Algorithm (GA) and Support Vector Machine (SVM) was developed for the study and Cleveland dataset from UCI machine learning library is used for the prediction. The prediction for coronary ailment was done using SVM and GA by optimizing hyper parameters of SVM: ‘C’ and ‘gamma’. The performance of heart disease classification is efficiently enhanced by implementing meta-heuristics and achieved 91% accuracy compare to SVM without GA. An approach of optimizing SVM parameters using GA outperforms SVM and SVM with k-cross validation for prediction heart diseases in terms of accuracy. It opens a direction to improve efficiency of machine learning algorithms.