Hybrid Feature Selection dan Ensemble Learning untuk Klasifikasi Risiko Stunting Anak di Indonesia

Abstract

Stunting is a chronic nutritional problem that remains a major public health issue in Indonesia. This study aims to develop a classification model for stunting risk in children using a combination of hybrid feature selection and ensemble learning methods. The dataset used is derived from socio-economic and health data obtained from the Central Statistics Agency and open datasets. The research method includes data preprocessing, feature selection, model development using Random Forest and Gradient Boosting combined with a Voting Classifier, and evaluation using accuracy, precision, recall, F1-score, and ROC-AUC metrics. The results show that the proposed model achieves high performance with accuracy reaching 98% and ROC-AUC close to 1. The hybrid feature selection successfully improves model efficiency by selecting relevant features. This study demonstrates that the integration of feature selection and ensemble learning can produce an accurate and interpretable model for early detection of stunting risk.