<![CDATA[Heart disease is one of the non contagious diseases that can lead to death. This disease occurs because of the narrowing of blood vessels that cause impairment of heart function. The death rate that caused by a heart disease is continuing increase and according by the Ministry of Health of the Republic of Indonesia research, in 2030 it reach 23.3 million peoples. It should be anticipated because the number of cardiologists in Indonesia is still very minimal. This research proposes framework Information Gain selection features with combination K-Nearest Neighbor and Naive Bayes to overcome the problems on the effectiveness and accuracy in classification heart disease. Information Gain algorithm used for reduce variable dimention to get relevant variables. After Information Gain selection features process is completed, the next process is classify numeric atributes with KNN and categorical atributes with Naive Bayes. The results of this research indicate an accuracy of 92.31% when the class distribution testing is balanced using 6 features with value of K=25 and when the class distribution testing is not balanced using 4 features with value of K=35. Based on these results, can be concluded that features selection Information Gain with combination KNN and Naive Bayes algorithm can be used for classifying heart disease.]]>
Copyrights © 2018
