MODEL PENGAMBILAN KEPUTUSAN DALAM PREDIKSI KASUS TUBERKULOSIS MENGGUNAKAN REGRESI LOGISTIK BERBASIS BACKWARD ELIMINATION

Abstract

The purpose of this study is to obtain a predictive model in cases of Tuberculosis (TB) using alogistic regression based on backward elimination. The study is conducted through literaturereview, collecting secondary data of 200 patients of Pulmonary Tuberculosis aged more than 15years.  Data is collected from BKPM Semarang, Central Java. The data consist ofcharacteristics, anamnesis, physical examination, laboratory test results data and radiologicalexamination. Logistic regression based  on backward elimination model has better accuracythan logistic regression without attribute selection.
Kata kunci : Tuberkulosis, Regresi Logistik, Backward Elimination, Prediksi