PERBANDINGAN KLASIFIKASI PENYAKIT HIPERTENSI MENGGUNAKAN REGRESI LOGISTIK BINER DAN ALGORITMA C4.5 (Studi Kasus UPT Puskesmas Ponjong I, Gunungkidul)

Abstract

Hypertension is a major problem in the world today. In Indonesia prevalence of hypertension is still high. There are two types of hypertension based on cause, primary and secondary hypertension. In this thesis focused on the classification of types of hypertension based on the cause using binary logistic regression and C4.5 algorithms with case studies in UPT Puskesmas Ponjong I, Gunungkidul of October-November 2015. Binary logistic regression is a method that describes the relationship between the response variable and several predictor variables with the variable equal to 1 to declare the existence of a characteristic and the value 0 to declare the absence of a characteristic. C4.5 algorithm is one method of classification of data mining is used to create a decision tree. The predictor variables were used in this thesis are gender, age, systolic blood pressure, diastolic blood pressure, treatment history, other diseases. To evaluate the result of classification use APER (Apparent Error Rate) calculation. Based on this analysis, classification of hypertension using binary logistic regression method obtained APER=27,4648% and 72,5352% of accuracy, while C4.5 algorithm obtained APER=35,9155% and 64,0845% of accuracy. In two different test proportion was found that there were significant differences of the two methods. Keywords : Types of Hypertension, Classification, C4.5 Algorithm, Binary Logistic Regression, APER