Ant colony optimization based support vector machine towards predicting coronary artery disease

Abstract

Data mining is the progression of finding the hidden information from the dataset available. Data mining started stepping and playing a major role in the medical field and it helps the medical practitioner to take a decision. Classification is considered as the major research issues in data mining. Classification is done based on the characteristics or features available. Currently, coronary artery disease (CAD) is getting evolve in South Asia countries, which is becoming a major cause for death. Data mining algorithms are being used to diagnosis of diseases, mostly in CAD. Currently available algorithms gets lack in classification towards accuracy. In this paper, a novel classification algorithm is proposed to effectively classify towards the prediction of CAD, namely ant colony optimization based support vector machine. It is designed to classify and predict CAD more accurately. The proposed algorithm classify the records in a dataset in a random manner instead of sequence manner. A threshold value is used for classification for more accurate results. The proposed algorithm is tested on Z-AlizadehSani dataset for the classification of heart disease among the patients. This research work uses the benchmark performance metrics namely sensitivity, specificity, and classification accuracy. The result shows that ACO-SVM is giving better results than SMO, BSMO, Bagging and NN algorithms.