CLASSIFICATION OF HEART DISABILITY BASED ON ELECTROCARDIOGRAM MEDICAL RECORDS USING DATA MINING CLASSIFICATION METHODS
Sumiati (a*), Donny Fernando (b), Ayu Purnama Sari (c), Felycia Felycia (d), Thoha Nurhadiyan H (e),Agung Triayudi (f)

a*,e, Informatic Department, Universitas Serang Raya, Indonesia
b,c, Information Systems Department, Universitas Serang Raya, Indonesia
d, Computer Systems Department, Universitas Serang Raya, Indonesia
f, Informatic Department, Universitas Nasional, Indonesia

Abstract

Diagnosing a patients heart disease is indeed not easy for a doctor, to diagnose a patient knowing a healthy or unhealthy heart is assisted by a tool, this tool is an electrocardiogram (EKG), besides that it requires the ability of a heart specialist who already has high abilities to be able to diagnose. accurately the patients type of heart defect based on the factors that exist in the patient. This study was conducted to test how accurate the results of the classification of heart disease were using electrocardiogram medical data. Initial research was conducted by comparing two classification methods, namely Naïve Bayes and K-Nearest Neighbor. The results of the comparison of the classification model show that the classification of heart defects using the K-Nearest Neighbor method has better accuracy than Naïve Bayes. System testing is carried out from the results of the validity of the system. The results of the accuracy with the Naïve Bayes method approach show a value of 75%, the result of the kappa value shows a value of 0.39%. While the accuracy results with the K-Nearest Neighbor approach show an accuracy value of 80%, and the kappa results show a value of 0.54.

Keywords: heart disease, electrocardiogram, classification, naïve bayes, accuracy , K-Nearest Neighbor

Topic: Symposium on Energy and Environmental Science and Engineering