Forecasting deep learning-based risk assessment of vector-borne diseases using hybrid methodology

Area of Science:

• Public Health
• Computational Biology
• Epidemiology

Background:

• Dengue fever is a growing concern in Malaysia, with cases nearly doubling in the past decade.
• Vector control is the primary strategy against dengue due to a lack of effective antiviral treatments.
• Vector-borne diseases pose a significant and increasing risk, causing substantial human illness.

Purpose of the Study:

• To propose a novel method for forecasting vector-borne disease risk.
• To enhance prediction accuracy using machine learning and optimization algorithms.
• To improve public health strategies for controlling vector-borne diseases.

Main Methods:

• Utilized Radial Basis Function Networks (RBFNs) for disease risk forecasting.
• Employed the Darts Game Optimizer (DGO) algorithm to enhance RBFN parameters.
• Trained models using historical disease data, climate variables, and geographical information.

Main Results:

• The RBFN-DGO model demonstrated superior predictive accuracy compared to standard methods.
• The DGO algorithm effectively optimized RBFN parameters for increased forecast precision.
• Extensive testing confirmed the robustness of the proposed RBFN-DGO approach.

Conclusions:

• The RBFN-DGO model offers a promising tool for predicting vector-borne disease risk.
• This research advances predictive capabilities in public health for disease control.
• The findings highlight the potential of optimized machine learning models in managing public health threats.