Related Experiment Video
Updated: Jan 8, 2026
08:20
Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
Published on: October 27, 2023
2.4K
Early Dengue Prediction in Bangladesh: A Comparative Study With Feature Analysis, Explainable Artificial
Md Atik Bhuiyan1, Md Rashik Shahriar Akash1, Radiful Islam1
1Department of Computer Science and Engineering, Daffodil International University, Dhaka, Bangladesh, daffodilvarsity.edu.bd.
Journal of Tropical Medicine
|December 16, 2025
Summary
A custom artificial neural network (ANN) achieved 97.5% accuracy in predicting dengue fever using patient symptoms, outperforming 12 other models. This highlights the potential of machine learning for early dengue detection.
Area of Science:
- Machine Learning
- Public Health
- Epidemiology
Background:
- Dengue fever is a significant public health concern in tropical and subtropical areas.
- Early detection of dengue is vital for effective management and control.
Purpose of the Study:
- To compare the effectiveness of 13 machine learning and deep learning models for nonclinical, symptom-based dengue prediction.
- To identify the most accurate model for dengue prediction in the Bangladeshi population.
Main Methods:
- A dataset of 500 patient records with 22 symptom-based features was used.
- Evaluated various classifiers including tree-based, linear, and instance-based models.
- A custom artificial neural network (ANN) was developed and hyperparameter-tuned.
Main Results:
- The hyperparameter-tuned ANN achieved the highest accuracy at 97.5%.
- Random forest models showed strong performance with 93.2% accuracy.
- SHapley Additive exPlanations (SHAP) identified key predictors like retro-ocular pain, swollen eyelids, and age.
Conclusions:
- A well-tuned ANN is superior for symptom-based dengue prediction.
- Broad model comparison and explainability are crucial for developing reliable diagnostic tools.
- Machine learning offers a promising approach for early dengue detection and public health intervention.
