Analyzing dengue outbreak patterns using integrated machine learning approaches: A study in Bangladesh.

Area of Science:

• Epidemiology
• Public Health
• Data Science

Background:

• Dengue fever is a significant global health concern, especially in tropical and subtropical regions.
• Existing challenges include data scarcity and complex transmission dynamics, hindering effective outbreak prediction.
• Early detection is crucial for implementing timely public health interventions.

Purpose of the Study:

• To develop and validate a hybrid machine learning framework for improved early detection and prediction of dengue fever outbreaks.
• To identify seasonal patterns and high-risk areas for dengue incidence.
• To address data limitations and transmission complexities using an integrated approach.

Main Methods:

• A hybrid machine learning methodology combining clustering and classification techniques was developed.
• Clustering analysis was performed on regional data to uncover latent patterns, with optimal clusters identified using silhouette scores.
• Supervised classification models were trained on meteorological and demographic data to predict dengue risk levels.

Main Results:

• Clustering analysis demonstrated robust data structures with a silhouette score of 0.774.
• Classification models achieved over 99% accuracy, precision, recall, and F1 scores.
• The models successfully identified high-risk periods and geographical regions with distinct seasonal dengue trends.

Conclusions:

• The developed data-driven framework enables proactive public health strategies for dengue outbreak management.
• This approach enhances the understanding of dengue transmission patterns.
• The methodology serves as a valuable tool for infectious disease control.