Using Machine Learning Algorithms to Determine the Food Insecurity Level of Households of Public School Children

• Main Authors: Cruz, Arnold Dela, Gallegos, Nathaniel Isaiah, Gattud, Kyle Aiden, Antonio, Victor A, Miro, Eden Delight, Go, Clark C
• Format: text
• Published: Archīum Ateneo 2024
• Subjects: Applied Mathematics; Food Security; Mathematics; Physical Sciences and Mathematics
• Online Access: https://archium.ateneo.edu/mathematics-faculty-pubs/251; https://doi.org/10.1063/5.0192150
• Institution: Ateneo De Manila University

Hunger and food insecurity continue to rise in the shadow of the COVID-19 pandemic affecting many vulnerable groups, especially children. As food is one of many fundamental human rights, looking into the problem contributes to helping uphold this basic right. Using survey data collected from households of public school children in a rural province and in a highly-urbanized city in the Philippines, we aim to compare three machine learning models, namely, logistic regression, support vector machine, and random forest, to predict the level of household food security based on geographic, household, and individual factors. A systematic assessment of the algorithms was performed by using accuracy, precision, recall, and F1-score, which showed that logistic regression algorithm performed best in predicting levels of food insecurity among Filipino households. This study shows that ML-based predictive models can potentially identify the food insecurity levels of a household, which can be used in improving the targeting mechanisms of nutrition-sensitive and nutrition-specific programs.