Modeling the Proportion of Tuberculosis Cases in South Sulawesi using Sparse Least Trimmed Squares

The deadliest infectious disease in Indonesia is tuberculosis (TB), and South Sulawesi is one of the provinces that contributed the most tuberculosis cases in Indonesia in 2018 with 84 cases per 100,000 population. This study aims to identify variables that could explain the proportion of TB cases i...

Full description

Saved in:
Bibliographic Details
Published inEKSAKTA: Journal of Sciences and Data Analysis
Main Authors Randa, Trigarcia Maleachi, Tinungki, Georgina Maria, Sunusi, Nurtiti
Format Journal Article
LanguageEnglish
Published 07.09.2022
Online AccessGet full text

Cover

More Information
Summary:The deadliest infectious disease in Indonesia is tuberculosis (TB), and South Sulawesi is one of the provinces that contributed the most tuberculosis cases in Indonesia in 2018 with 84 cases per 100,000 population. This study aims to identify variables that could explain the proportion of TB cases in South Sulawesi. The data used has many explanatory variables, and there are outliers. Sparse Least Trimmed Squares (LTS) analysis can be used to handle data that has many explanatory variables and outliers. The resulting sparse LTS model successfully selects and shrinks the variables to 14 variables only. In addition, based on the value of R2 and RMSE for the model evaluation, the sparse LTS shows satisfying results rather than classical LASSO. The government can focus on these factors if they want to reduce the proportion of TB cases in South Sulawesi.
ISSN:2720-9326
2716-0459
DOI:10.20885/EKSAKTA.vol3.iss2.art6