Predicting Global Youth Literacy Rates using Machine Learning: A Feature Engineering Approach

Youth literacy remains a critical global challenge despite notable progress. This study employs machine learning to predict youth literacy rates (ages 15–24) across countries using a global education dataset. A novel feature— the Overall Education Completion Index (OECI)—was engineered to summarize educational attainment and improve model accuracy. Multiple classifiers were evaluated, with Random Forest achieving ~93.7% accuracy, meeting the target range of 89–97%. Key predictors include birth rate, tertiary enrollment, and out-of-school rates, highlighting links between education and demographic trends. Feature importance analysis revealed birth rate as the strongest predictor, followed by tertiary education. The model’s insights support data-driven policy planning and monitoring of SDG 4 literacy goals. Visualization tools such as correlation heatmaps and confusion matrices validate model interpretability. Results affirm the power of thoughtful feature engineering in boosting predictive performance. This study provides a scalable, interpretable framework for global education monitoring using machine learning.