Classification Of Stunting In Toddlers Using The Random Forest Method

Sutarti Dwi Yanti (a:1:{s:5:"en_US"s:37:"Muhammadiyah University of Pontianak "})
Menur Wahyu Pangestika ([e-mail])
Alda Cendekia Siregar ([e-mail])

International Journal of Health, Engineering and Technology
Vol. 5 No. 1 (2026): IJHET MAY 2026

Monitoring and data collection related to stunting at several community health centers (Puskesmas) in Ketapang Regency play a crucial role in assessing optimal growth and development of fetuses and newborns. One of the recurring issues in Ketapang Regency is the inaccuracy and inconsistency in monthly stunting data collection. This study aims to design a stunting classification model for children under five using the random forest method and to evaluate its classification accuracy. The research follows several stages problem identification, literature review, data collection, data processing, testing, and drawing conclusions. The performance of the random forest method is assessed to determine the impact of each stage on the model’s classification ability. Evaluation metrics are derived from the confusion matrix. The confusion matrix results show a recall of 97%, precision of 96%, F1-score of 96%, and an accuracy of 91%, indicating that the random forest method performs excellently in classifying nutritional status.