Skip to content
SDGs Center UDINUS

Evidence record · UDINUS · SDGs

Applying Machine Learning to Analyze Tuberculosis Service Standards in Indonesia

Rektorat UDINUS 2026 Published 2026-10-30

Evidence record · UDINUS · SDGs

Unit
Rektorat UDINUS
Period
2026
Goals
Record
c7637285-7388-4649-bcfb-cb760f5a5fbb
Published
2026-10-30

Universitas Dian Nuswantoro conducted a study applying machine learning to analyze tuberculosis service standards in Indonesia. The activity involved Catur Yuantari, Maria G.; Manglapy, Yusthin M.; Nurmandhani, Ririn; Jaya Kusuma, Edi; Haikal; and Miarso, Dani as participants. It addressed tuberculosis as a significant public health challenge in Indonesia, where disparities in service quality persist. The activity focused on early identification of substandard tuberculosis services to help prevent transmission and accelerate elimination efforts.

The study used a retrospective design and analyzed primary and secondary data from tuberculosis patients in Semarang City, Indonesia. Random Forest, K-Nearest Neighbors, and Boosting algorithms were implemented. The analytical process included stratified cross-validation, feature selection through Recursive Feature Elimination, hyperparameter tuning, and variable importance analysis. Model performance was evaluated using accuracy, area under the curve, recall, and F1-score, with particular emphasis on detecting the minority class of non-compliant services.

The Random Forest model achieved the highest reported test accuracy, with an average area under the curve also reported. These results were interpreted cautiously because the class distribution was highly imbalanced, limiting the model's ability to detect minority cases. The analysis therefore emphasized recall and F1-score for the minority class as indicators of practical usefulness for early detection. Five key predictors were identified, with anti-tuberculosis drug regimen and age contributing most to model performance. The model showed potential as an early warning tool for substandard tuberculosis services and could support risk-based interventions. The activity provides evidence to guide targeted tuberculosis control policies and may support decision-making related to tuberculosis elimination.

Reference https://www.scopus.com/pages/publications/105040509173?origin=resultslist