Early Detection of Mental Health Risk Indicators in Children Using Machine Learning Based on Teacher Questionnaires in Islamic Early Childhood Education in Gorontalo.

Abstract

This study aims to identify indicators of mental health risk in early childhood through the development of a machine learning-based system and to analyze its implications for education. Mental health in early childhood is a crucial aspect that supports optimal development. Various internal and external factors influence the development of children's mental health. Early detection of risk indicators enables appropriate interventions to prevent more serious problems in the future. This research utilizes data collected from 100 randomly selected PIAUD (Islamic Early Childhood Education) teachers in Gorontalo Province. The K-Means Clustering algorithm is used to group the data and form target variables, while the Decision Tree algorithm is employed for classification. The results show that the Decision Tree model achieves an accuracy of 85% in predicting mental health risks in children. Indicators such as “Withdrawal,” “Easily Angered,” “Weight Change,” and “Eating Problems” are identified as key factors. This prediction system is expected to serve as a helpful tool for teachers and parents in conducting early detection and providing appropriate interventions.