Svm Based Decision Support System for Heart Disease Classification with Integer-coded Genetic Algorithm to Select Critical Features

The paper presents a decision support system for heart diseases classification based on support vector machines (SVM) and integer-coded genetic algorithm (GA). Simple Support Vector Machine (SSVM) algorithm has been used to determine the support vectors in a fast, iterative manner. For selecting the important and relevant features and discarding those irrelevant and redundant ones, integer-coded genetic algorithm is used which also maximizes SVM‘s classification accuracy. The heart disease database used in this study includes 303 cases, and 13 diagnostic features were used for each case. The results of the 5-class classification problem indicate an increase in the overall accuracy using the optimal feature subset, accuracy achieved being 72.55% indicating the potential of the system to be used as a practical decision support system. As a two class problem, the proposed method gives an accuracy of 90.57% which is better than the existing methods.