Penerapan Metode Exhaustive Chi-Square Automatic Interaction Detection pada Klasifikasi Penderita Diabetes dan Non Diabetes

Abstract

Classification is a process of grouping an object into a certain category. One of classification method is the Exhaustive Chi-Square Automatic Interaction Detection (CHAID). The Exhaustive CHAID method is a classification method for categorical data by forming a classification tree. The classification tree interprets predictor variables that have a significant effect on the response variable based on the chi-square test. The purpose of this study was to obtain classification results for diabetics and non-diabetics using the Exhaustive CHAID method. The response variable used is the blood sugar level and the predictor variables consist of systolic blood pressure, diastolic blood pressure, length of sleep, working style, level of knowledge about diabetes, abdominal circumference, hereditary history of diabetes, age, exercise habits, and body mass index. The classification results show that the factors that have a significant influence at the 5% level are a hereditary history of diabetes, abdominal circumference, level of knowledge about diabetes, and diastolic blood pressure. Apart from that, the accuracy value of the Exhaustive CHAID classification tree is quite good, namely 86% based on the confusion matrix.