A MACHINE LEARNING FRAMEWORK FOR REGIONAL LIFE EXPECTANCY PREDICTION

Abstract

Life expectancy is a crucial indicator of a population's overall health and standard of living. This work offers an AI-driven approach for accurately predicting life expectancy using state-of-the-art machine learning algorithms. In particular, the CatBoost approach—which is renowned for its superior performance on structured data and categorical variables—was used to build a robust prediction model. The model was trained using a big dataset of health, demographic, and socioeconomic data from multiple countries. Performance evaluation results were positive, with a Mean Squared Error (MSE) of 0.18, Mean Absolute Error (MAE) of 1.15, testing R2 score of 93%, and training R2 score of 95%. These metrics confirm the model's excellent generalisation to new data and strong prediction power. By identifying key factors that affect life expectancy, this AI-powered approach can support health policy planning, improve resource allocation, and promote preventative healthcare initiatives across communities. The project shows how machine learning can be effectively used to provide insightful information on longevity prediction and public health. Keywords: Machine Learning, CatBoost, Life Expectancy Prediction, Socioeconomic Factors, HealthCare data and Regression Models.