Version 1
: Received: 18 September 2024 / Approved: 19 September 2024 / Online: 19 September 2024 (13:18:24 CEST)
Version 2
: Received: 23 September 2024 / Approved: 23 September 2024 / Online: 23 September 2024 (14:31:09 CEST)
Version 3
: Received: 23 September 2024 / Approved: 23 September 2024 / Online: 23 September 2024 (14:39:40 CEST)
How to cite:
Anh Tuan, D.; Dang, T. N. Integrating Climate Data and Advanced Machine Learning for Precision Dengue Outbreak Prediction: A Study in Ba Ria-Vung Tau province, Vietnam. Preprints2024, 2024091535. https://doi.org/10.20944/preprints202409.1535.v2
Anh Tuan, D.; Dang, T. N. Integrating Climate Data and Advanced Machine Learning for Precision Dengue Outbreak Prediction: A Study in Ba Ria-Vung Tau province, Vietnam. Preprints 2024, 2024091535. https://doi.org/10.20944/preprints202409.1535.v2
Anh Tuan, D.; Dang, T. N. Integrating Climate Data and Advanced Machine Learning for Precision Dengue Outbreak Prediction: A Study in Ba Ria-Vung Tau province, Vietnam. Preprints2024, 2024091535. https://doi.org/10.20944/preprints202409.1535.v2
APA Style
Anh Tuan, D., & Dang, T. N. (2024). Integrating Climate Data and Advanced Machine Learning for Precision Dengue Outbreak Prediction: A Study in Ba Ria-Vung Tau province, Vietnam. Preprints. https://doi.org/10.20944/preprints202409.1535.v2
Chicago/Turabian Style
Anh Tuan, D. and Tran Ngoc Dang. 2024 "Integrating Climate Data and Advanced Machine Learning for Precision Dengue Outbreak Prediction: A Study in Ba Ria-Vung Tau province, Vietnam" Preprints. https://doi.org/10.20944/preprints202409.1535.v2
Abstract
Dengue fever is a persistent public health issue in tropical regions, including Vietnam, where climate variability significantly influences transmission dynamics. This study aims to develop machine learning models to forecast dengue outbreaks in Ba Ria-Vung Tau province, Vietnam, by leveraging meteorological data from 2003 to 2022. Four models were utilized: Negative Binomial Regression (NBR), Seasonal AutoRegressive Integrated Moving Average with Exogenous Regressors (SARIMAX), Extreme Gradient Boosting (XGBoost), and Long Short-Term Memory (LSTM) networks. Key climate variables incorporated in the models include daily maximum and minimum temperature (ranging from 26.57°C to 29.63°C), temperature range (1.86°C to 7.26°C), relative humidity (58.3% to 89.1%), precipitation (up to 42.53 mm/day), surface pressure (100.08 to 101.14 kPa), wind speed (1.90 to 10.23 m/s), wind direction, and sea surface temperature (24.7°C to 29.8°C). Lagged variables from 2 to 20 weeks were included to account for delayed climatic effects on dengue transmission. The NBR model demonstrated the highest predictive accuracy with the lowest Mean Absolute Error (MAE) of 21.41. SARIMAX and LSTM models effectively captured seasonal trends but struggled with short-term outbreak prediction, achieving MAEs of 20.31 and 28.86, respectively. XGBoost exhibited moderate predictive performance (MAE: 24.45) but was prone to overfitting without fine-tuning. These findings highlight the value of climate-based machine learning models, particularly NBR, in forecasting dengue outbreaks in Ba Ria-Vung Tau. However, enhancing short-term outbreak prediction remains challenging, underscoring the need for model refinement and integration into early warning systems for more effective public health responses.
Public Health and Healthcare, Health Policy and Services
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
