ARTIFICIAL INTELLIGENCE–DRIVEN SPATIAL PREDICTION OF CLIMATE-SENSITIVE INFECTIOUS DISEASE HOTSPOTS IN PAKISTAN

Abstract

Climate change is altering the spatial and temporal dynamics of infectious diseases through changes in temperature, precipitation, humidity, ecological conditions, and vector suitability. Pakistan is particularly vulnerable because of climatic diversity, rapid urbanization, population density, recurrent flooding, and uneven public-health infrastructure. This study proposes an artificial intelligence (AI)-driven geospatial framework for predicting climate-sensitive infectious-disease hotspots in Pakistan. District-level epidemiological records are integrated with climatic, environmental, demographic, geographic, and remotely sensed variables. Random Forest, XGBoost, Support Vector Machine, and Artificial Neural Network models are compared using temporal and spatial validation. Explainable-AI techniques are used to identify influential predictors, and the best-performing model is used to generate district-level risk maps. The framework is designed to capture nonlinear and interacting relationships and to support climate-informed surveillance, targeted prevention, and risk-based public-health resource allocation. Because the source manuscript does not contain empirical model outputs, numerical findings are not fabricated; the proposed analytical and interpretive framework is retained for completion after dataset analysis.