IMPLEMENTASI MODEL REGRESI LOGISTIK MULTINOMIAL PADA PENGELOMPOKAN PENYAKIT TUBERKULOSIS

Abstract

There are three types of Tuberculosis (TB) patients at Banyumas Region Hospital, namely negative pulmonary TB, positive pulmonary TB, and extra pulmonary TB. Types of TB generally caused by age, cae of history, gender, level of education, and domicile. One of the methods that used to find a correlation between types of TB with the affect is regression analysis. This study used multinomial logistic regession analysis because types of TB is categorical and the data is 156 TB’s patients recorded at 2018/2019. The result showed that the level of education be a dominant factor to affect TB. Here, we noted that patients with basic education level have a 5,843 time odds for getting positive pulmonary TB and 2,224 times for getting extra pulmonary TB. The multinomial logistic regression model is then given as probability for getting positive pulmonary TB with factor level of education is greather than negative pulmonary TB and extra pulmonary TB.