Machine Learning-Based Prediction of Heatwave-Related Hospitalizations: A Case Study in Matam, Senegal
2025
Mory Toure | Ibrahima Sy | Ibrahima Diouf | Ousmane Gueye | Endalkachew Bekele | Md Abul Ehsan Bhuiyan | Marie Jeanne Sambou | Papa Ngor Ndiaye | Wassila Mamadou Thiaw | Daouda Badiane | Aida Diongue-Niang | Amadou Thierno Gaye | Ousmane Ndiaye | Adama Faye
This study assesses the impact of heatwaves on hospital admissions in the Matam region of Senegal by combining climatic indices with machine learning methods. Using daily maximum temperature (TMAX) and heat index (HI), heatwave events were identified from 2017 to 2022. Hospital data from Ourossogui Regional Hospital were analyzed, and three predictive models, Random Forest (RF), Extreme Gradient Boosting (XGB), and Generalized Additive Models (GAMs), were compared. A bootstrapping approach with 1000 iterations was used to evaluate model robustness. The findings reveal a significant delayed effect of heatwaves, with increased hospitalizations occurring three to five days after the event. RF outperformed the other models with R2 values ranging from 0.51 to 0.72. These findings highlight the need to enhance heatwave monitoring and promote the integration of impact-based climate forecasting into health early warning systems, particularly to protect vulnerable groups such as the elderly, children, and outdoor workers.
Показать больше [+] Меньше [-]Ключевые слова АГРОВОК
Библиографическая информация
Эту запись предоставил Multidisciplinary Digital Publishing Institute