Identifying Risk Factors and Predicting Food Security Status Using Supervised Machine Learning Algorithms

Abstract

This project focuses on leveraging various super- vised machine learning techniques to analyse and predict food security status using historical data. It aims to identify key risk factors and provide predictive models that stakeholders can use to anticipate food insecurity in different populations. In addition to enhancing the decision-making process, the project also seeks to support early warning systems that can alert authorities about potential food crises. By analysing diverse variables such as household demographics, agricultural production, climatic conditions, and economic indicators, the model can offer valuable insights into regional disparities and vulnerability patterns. The models developed in this project can be trained using datasets collected from international organizations such as the World Bank, FAO, and WFP. These datasets may include metrics like food consumption scores, market access, income levels, rain- fall distribution, and crop yields. Furthermore, the integration of geospatial data and time series analysis can strengthen the accuracy and timeliness of predictions. A range of supervised learning algorithms will be explored and compared, including Logistic Regression, Decision Trees, Random Forests, Support Vector Machines (SVM), and Gradient Boosting techniques such as XG Boost and Light GBM. The project will also include rigorous preprocessing steps such as handling missing values, encoding categorical variables, normal- ization, and feature selection to enhance model performance.

Introduction

Global food security is increasingly affected by fluctuating food availability, resource constraints, and environmental instability. Traditional survey-based monitoring is often slow, resource-intensive, and unable to capture rapid changes. To address this, the Food Security Prediction System applies supervised machine learning—including XGBoost, Random Forest, Logistic Regression, and ensemble models—to predict regional and household-level food security using socio-economic, agricultural, and climatic data.

The system integrates multiple intelligent modules:

• Regional Vulnerability Assessment – identifies high-risk areas and seasonal trends.
• Geospatial Visualization – interactive heatmaps and district-wise risk maps for real-time insights.
• Model Explainability – feature importance charts, confusion matrices, and transparency tools.
• Hybrid Predictive Engine – combines multiple ML models to improve accuracy.
• Resource Planning and Alert System – supports proactive intervention, food allocation, and policy decisions.

The platform is deployed as a lightweight, web-based interface using Python and Streamlit, enabling accessible, interactive dashboards for NGOs, government agencies, and humanitarian planners. By combining predictive analytics, explainable AI, and geospatial visualization, the system supports data-driven decision-making, proactive planning, and efficient food security management, with future enhancements including real-time data integration and automated alerts.

Conclusion

The Food Security Prediction System developed in this project demonstrates the potential of Artificial Intelligence (AI) and Machine Learning (ML) to support governmentagencies, NGOs, and policymakers in proactively addressing food insecurity. By analyzing historical agricultural, climatic, economic, and satellite-derived features, the system accurately predicts regional food security status and identifies locations at risk of food shortages.

References

[1] S. P. Fraiberger, D. B. Reeves, B. Cesare, E. Chen, and S. Krishnan, ”Tracking Global Food Insecurity using News Media: A Deep Learning Approach,” in npj Digital Medicine, vol. 5, no. 65, pp. 1–10, 2022. [2] A. Kamilaris and F. X. Prenafeta-Boldu´, ”Deep Learning in Agriculture: A Survey,” in Computers and Electronics in Agriculture, vol. 147, pp. 70–90, 2018. [3] D. Bertetti, A. Sonesson, and M. Luccioni, ”Predicting Food Crises by Integrating Conflict Data and Environmental Variables,” in Proc. ACM Conference on AI for Good, pp. 1–9, 2023. [4] S. Nosratabadi, M. Mosavi, N. Shamshirband, A. Zavadskas, A. Rako- tonirainy, and F. Jolai, ”Crop Yield Prediction with Hybrid Machine Learning Models,” in Sustainability, vol. 11, no. 1, pp. 1–20, 2019. [5] Y. Xie, X. Jean, M. Burke, D. Lobell, and S. Ermon, ”Transfer Learning from Deep Features of Nighttime Lights for Poverty Mapping,” in Proceedings of the National Academy of Sciences (PNAS), vol. 113, no. 52, pp. E7846–E7853, 2016. [6] M. Nazari, S. Kahaki, and B. Mahmoodian, ”Modeling Agricultural and Livestock Production using MLP and ANFIS,” in Information Processing in Agriculture, vol. 7, no. 2, pp. 321–336, 2020. [7] T. Nigus and H. Shashirekha, ”Household Food Security Prediction using Ensemble Learning with Feature Elimination,” in Proc. IEEE International Conference on Data Science and Knowledge Engineering (ICDSKE), pp. 1–6, 2022. [8] Y. Liu, L. Zhang, and H. Sun, ”Predicting Rural Household Food Insecurity in China using Random Forest and LASSO Regression,” in Journal of Rural Studies, vol. 103, pp. 10–21, 2023. [9] E. Knippenberg, N. Jensen, and L. Constas, ”Forecasting Food Insecurity using Machine Learning and Household Surveys,” in World Develop- ment, vol. 159, pp. 106–120, 2022. [10] R. Mitra and S. Ackermann, ”Machine Learning Forecasting Models for Food Insecurity in Malawi: A Case Study,” in Proc. IEEE International Conference on Big Data, pp. 1520–1527, 2023. [11] J. Silva, M. Costa, and R. de Oliveira, ”Predicting Food Insecurity using Socioeconomic Indicators: A Machine Learning Approach,” in Sustainability, vol. 14, no. 3, pp. 1–12, 2022. [12] P. Sood and D. Singh, ”A Comprehensive Review on Machine Learning and Computer Vision Applications to Food Security,” in Artificial Intelligence in Agriculture, vol. 6, pp. 10–28, 2022.

Copyright

Copyright © 2026 Ms. Y Sujana, A. Yasaswini, P. Vinuthna, V. Praveen. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.