Machine Learning with Neural Networks and Random Forest to Predict Nutritional Risk in Children Under 5 Years of Age

Child malnutrition has a high prevalence and is associated with multiple risk factors. Although established treatments and methodologies exist, there is still a need for innovative approaches that enable early detection and more effective intervention. This study implemented a predictive model based on machine learning (ML), using algorithms such as neural networks and random forest. The Knowledge Discovery in Databases (KDD) methodology was applied to structure the analysis process and ensure accurate results. A dataset of 500 randomly selected patients was used, including anthropometric, clinical, and socioeconomic variables. The developed model effectively identified the presence of child malnutrition and the key factors associated with the condition. The analysis revealed that 89% of the cases presented some degree of malnutrition, indicating a significant risk pattern in the studied population. The application of predictive models based on advanced ML techniques offers an effective tool for the early identification of child malnutrition, potentially improving prevention and treatment strategies and strengthening the response to this critical public health issue.