A Geospatial Analysis of Food Insecurity Among Refugee Households in Lebanon Using Machine Learning Techniques

This study integrates geospatial analysis with machine learning to understand the interplay and spatial dependencies among various indicators of food insecurity. Combining household survey data and novel geospatial data on Syrian refugees in Lebanon, we explore why certain food security measures are effective in specific contexts while others are not. Our findings indicate that geolocational indicators significantly influence food insecurity, often overshadowing traditional factors like household socio-demographics and living conditions. This suggests a shift in focus from labor-intensive socioeconomic surveys to readily accessible geospatial data. The study also highlights the variability of food insecurity across different locations and subpopulations, challenging the effectiveness of individual measures like FCS, HDDS, and rCSI in capturing localized needs. By disaggregating the dimensions of food insecurity and understanding their distribution, humanitarian and development organizations can better tailor strategies, directing resources to areas where refugees face the most severe food challenges. From a policy perspective, our insights call for a refined approach that improves the predictive power of food insecurity models, aiding organizations in efficiently targeting interventions.