Centroid neural networks for clustering nations: A machine learning approach to humanitarian aid allocation

HELP International, a leading humanitarian NGO focused on alleviating poverty and responding to global crises, is tasked with the effective allocation of a substantial $10 million aid package. To support this mission, this study applies unsupervised machine learning techniques to group countries according to a diverse set of socio-economic and health indicators. Key features used in this analysis include child mortality, exports, imports, income, GDP per capita, health expenditure, inflation, life expectancy, and total fertility rate. Data preprocessing and normalization were conducted to ensure data integrity and comparability, followed by Principal Component Analysis (PCA), which successfully retained over 90% of the dataset’s variance while reducing dimensionality. To determine the optimal number of clusters, the Elbow method was employed. A Centroid Neural Network (CNN) algorithm was then applied, outperforming traditional clustering techniques, as demonstrated by favorable silhouette coefficients, Calinski–Harabasz indices, and Davies–Bouldin scores. The resulting clusters reveal clear distinctions between countries with robust socio-economic conditions and those facing critical challenges such as high child mortality, low income levels, and severe inflation. The analysis culminates in a visually intuitive choropleth map that spatially represents these clusters, providing HELP International with a powerful, data-driven decision-making tool. This methodology ensures that limited financial resources are directed to the countries in greatest need, enhancing the impact of humanitarian interventions and promoting long-term development. The findings not only serve as a practical guide for HELP International’s current funding priorities but also establish a replicable framework for future aid distribution efforts. By grounding allocation strategies in objective, data-driven insights, this study contributes to more equitable and effective global resource deployment.