PEMODELAN FAKTOR-FAKTOR YANG MEMENGARUHI KESEMBUHAN TUBERKULOSIS DI PROVINSI JAWA TIMUR DENGAN REGRESI NONPARAMETRIK SPLINE TRUNCATED

Abstract

Tuberculosis (TB) is a disease caused by Mycobacterium tuberculosis. TB transmission can be through environmental conditions (air) and community behavior. East Java Province has 64,764 cases of tuberculosis. The process of curing TB disease is essential to reduce the number of cases or break the chain of transmission. Many factors are thought to influence the tuberculosis cure rate, so an analysis was carried out that aims to model the factors that have a significant effect. This study consisted of six predictor variables, including food management places meeting health requirements, households having proper sanitation, complete tuberculosis treatment, public places  meeting health requirements, health insurance, and smoking residents over 15 years. The method used in modeling the tuberculosis cure rate in East Java in 2020 is nonparametric spline truncated linear regression with a combination of knot points (2,3,1,3,2,3). The optimal knot point selection was obtained from the lowest Generalized Cross Validation (GCV) value of 30.7126. This results of study obtained the value of the adjusted coefficient of determination 82.46 percent and Mean Square Error (MSE) of 15.4028, which indicates that the linear truncated spline regression model with a combination of knot points is feasible or good to use in modeling the tuberculosis cure rate in East Java in 2020.

Keywords: Spline truncated, tuberculosis cure, knots, Generalized Cross Validation (GCV)