RETRACTED ARTICLE: A hybrid approach for mortality prediction for heart patients using ACO-HKNN

Abstract

Heart disease is the major cause of mortality in the world. The heart disease prediction from the clinical data is deliberate as the most important subject in clinical data analysis. Especially the size of data in health care is vast. Data mining (DM) assists decision and prediction from the raw health care data. DM converts the large collection into useful information. Several existing studies utilize the data mining approaches in heart disease prediction. There is only little research focused on selecting the important features which play a significant role in predicting heart disease is less. The aim of this study is to provide an enhanced approach with novel feature selection and classification technique to predict mortality in congestive heart failure patients. Through this approach the death rate due to heart disease will be decreased gradually. The ant colony optimization (ACO) algorithm is utilized for selecting the best feature for hybrid K-nearest neighbor (KNN) classifier. The proposed approach is compared with the prior classification techniques such as the Support vector machine, Naive Bayes, KNN, C4.5, and decision tree. UCI Cleveland dataset is utilized for our implementation. Using the Netbeans IDE an experimental was conducted and the result shows that the heart disease prediction model provides a better result with accuracy of 99.2%.The present study shows the efficiency of the HKNN in heart disease prediction system. Initially important features are analyzed and then classification is utilized to obtain a better result.