Klasifikasi Kegagalan Pengobatan Penyakit Tuberkulosis (TB) di Kota Mataram Menggunakan Metode Multivariate Adaptive Regression Splines (MARS)

Abstract

Tuberculosis (TB) is an infectious disease that affects the lungs and can spread through the air. Indonesia ranks second globally, contributing approximately 10% of the total TB cases worldwide. The Province of Nusa Tenggara Barat (NTB) has been recorded as having a high number of TB cases, with the City of Mataram being one of the areas with the highest incidence. Although TB can be cured through anti-tuberculosis drug (OAT) treatment for six months, the success rate of this treatment has declined since 2016. Several factors such as age, gender, education level, and other health conditions may influence treatment success. Therefore, evaluating treatment outcomes is very important, including monitoring the results of treatment to determine whether the treatment is successful or unsuccessful. In order to classify TB treatment failure, an effective statistical method that can be used is Multivariate Adaptive Regression Splines (MARS). MARS is a flexible nonparametric regression method capable of handling high-dimensional data, making it very useful for classifying data with many predictor variables. This study aims to classify TB treatment failure in the City of Mataram using the MARS method, with the expectation of improving treatment success in the region. Based on the analysis using the MARS method, the type of TB (X₅) was found to be the main determining variable of treatment outcomes, with a fairly good overall classification accuracy of 80%.