MODEL DIAGNOSIS TUBERKULOSIS MENGGUNAKAN k-NEAREST NEIGHBOR BERBASIS SELEKSI ATRIBUT

Abstract

The objective of this paper is to obtain a diagnosis model of Tuberculosis (TB) using k-Nearest Neighbor based on feature selection. Data is collected from BKPM Semarang, Central Java. The data consist of characteristics, anamnesis, physical examination, laboratory test results, radiological examination, duration of cough and sputum color. The results indicate that the k-Nearest Neighbor based on backward elimination model improvements as high as 78.66% % compared to individual models.
Keywords: k-Nearest Neighbor, backward ellimination , Tuberkulosis, diagnosis, pengambilan keputusan