Deteksi Dini Stunting pada Balita Menggunakan Data Mining dengan Algoritma C4.5

Abstract

Stunting is a serious issue affecting the growth and development of children in Indonesia, with a prevalence still high, reaching 148 million children under five. This study aims to develop an early detection model for stunting using the C4.5 decision tree algorithm, utilizing a large dataset containing 120,999 records that include attributes of age, height, and gender. The research method used is a quantitative experimental approach with data mining techniques, where the model was evaluated using 10-fold cross-validation to ensure accuracy and generalizability. The results show that the C4.5 model achieves 99.87% accuracy, with very high precision and recall, and good interpretability, making it suitable for implementation in public health systems. These findings emphasize the importance of height as a key indicator in detecting stunting and provide a basis for model integration in digital health initiatives in Indonesia. This study recommends incorporating socioeconomic and environmental attributes for more comprehensive analysis in the future.