DEVELOPMENT OF DATA-DRIVEN SYSTEM FOR EARLIER CHILDHOOD MALNUTRITION PREDICTION

Abstract

Childhood malnutrition remains a critical issue in Tanzania, impacting the health and development of children under five. Early detection and intervention are vital to mitigating its effects, yet they are often hampered by a lack of effective tools. This study addresses this challenge by developing a data-driven system that uses machine learning to predict malnutrition risks early in childhood. Unlike previous studies that focused on limited data types or specific regions, this research presents an original approach that integrates multiple data categories to enhance prediction accuracy and relevance to the Tanzanian context. T he system analyzes a range of factors, including socioeconomic factors (poorest, Urban-Rural), health data such (height, weight, stunted, wasted, underweight, sex and age) and environmental variables (healthy status), to identify at-risk children before they exhibit significant symptoms. By leveraging a Random Forest algorithm, the study achieved a high accuracy of 96%, demonstrating the model's strong predictive performance. The data used for model development were obtained from a publicly available Kaggle dataset, which provides a valuable foundation but also represents limitations, as the secondary and non-Tanzanian data may affect the model’s generalizability to local contexts.