Related Experiment Video
Updated: Jul 11, 2025
10:35
Vector Competence Analyses on Aedes aegypti Mosquitoes using Zika Virus
Published on: May 31, 2020
3.0K
Predicting dengue transmission rates by comparing different machine learning models with vector indices and
Song Quan Ong1, Pradeep Isawasan2, Ahmad Mohiddin Mohd Ngesom3
1Entomology Laboratory, Institute for Tropical Biology and Conservation, Universiti Malaysia Sabah, Jalan UMS, 88400, Kota Kinabalu, Sabah, Malaysia. songquan.ong@ums.edu.my.
Scientific Reports
|November 5, 2023
Summary
Machine learning models can predict dengue transmission using weather and vector data. Ensemble methods like XGBoost show superior performance, improving early warning systems.
Area of Science:
- Epidemiology
- Data Science
Background:
- Machine learning (ML) is increasingly used for dengue transmission prediction.
- Existing models often use limited variables and algorithms, necessitating improved approaches.
Purpose of the Study:
- To develop and compare ML models using diverse predictors for enhanced dengue transmission monitoring.
- To identify high-performing algorithms and key predictive variables.
Main Methods:
- Trained and validated seven ML algorithms, including ensemble methods, using vector indices and meteorological data.
- Evaluated model performance using ROC AUC, accuracy, and F1 score.
- Assessed variable importance and impact of feature removal.
Main Results:
- Ensemble ML methods (XGBoost, AdaBoost, Random Forest) outperformed traditional algorithms.
- XGBoost achieved the highest AUC, accuracy, and F1 score.
- Removing the container index improved model performance by at least 6%.
Conclusions:
- Ensemble ML models offer a robust framework for dengue transmission prediction.
- Variable selection, excluding the container index, enhances predictive accuracy.
- This study provides a foundation for developing effective dengue early warning systems.
