Heart Disease Classification Model Using K-Nearest Neighbor Algorithm

Abstract

Heart disease is a disease that needs to be watched out for and is of particular concern. Seeing to the WHO report, in 2018, as many as 17.9 million people died from heart disease, and especially in Indonesia, heart disease in 2020 became the highest cause of death. This study uses data mining techniques to pull out information from the data used. This research provides a scientific contribution, namely detecting heart disease as early as possible. In this case, the author uses the K-Nearest Neighbor Algorithm to classify the data based on the nearest neighbor data. The database is own in a reasonably high volume, so it should note that irrelevant attributes will be removed over or noise. If they are still used, data processing results will not be optimal, so data cleaning needs to be done carefully. The selection of the data used was 1243 records, and after being selected the data were taken in this study as many as 366 records, with parameters using 12 attributes, actual data from hospitals, data consisting of data from patients under surveillance for cardiac care, and data from patients who underwent surgery and Data from Medical Examination. Therefore, it is necessary to develop a decision support system that assists doctors in taking steps for early detection. Research conducted with the K-Nearest Neighbors algorithm accuracy up to 77% with a value of K = 7.