Probabilistic and Interpretable Approach for Malaria Transmission Forecasting via ARIMA

The current Malaria outbreak poses a significant global challenge for healthcare systems worldwide. Each day, new and intricate datasets emerge, detailing positive and negative case counts, hospital admissions, mortality rates, bed occupancy, ventilator shortages, and more. Infections surge notably during the rainy season, underscoring the urgent need for early detection and forecasting techniques to curb disease transmission and facilitate a swift and safe resumption of economic activities. A model was devised utilizing classical mathematical modeling methods, employing exponential regression, to forecast infection spread. The accuracy of these forecasts was assessed, revealing limitations in predicting infection spread beyond a two-week timeframe. Results underscore the efficacy of the ARIMA model for short-term forecasting, enabling timely public health responses and evidence-based interventions crucial for effective disease management and containment. Leveraging automated model selection processes and precise performance evaluation metrics enhances the reliability and applicability of forecasting tools in real-world healthcare scenarios, aiding decision-making and resource allocation strategies.