[Introduction on a forecasting model for infectious disease incidence rate based on radial basis function network].

Area of Science:

• Epidemiology
• Computational Biology
• Artificial Intelligence

Context:

• Accurate infectious disease forecasting is vital for effective prevention and control strategies.
• Traditional mathematical models often struggle with the complex, nonlinear factors influencing disease incidence.
• The need for advanced predictive tools is increasing due to the dynamic nature of infectious diseases.

Purpose:

• To introduce and evaluate the efficacy of a Radial Basis Function (RBF) neural network for predicting infectious disease incidence rates.
• To compare the performance of the RBF network model against the Autoregressive Integrated Moving Average (ARIMA) model.
• To assess the reliability of the RBF model using historical hepatitis B incidence data.

Summary:

• A forecasting model utilizing Radial Basis Function (RBF) neural networks was developed using hepatitis B monthly incidence data from 1991-2002.
• The model was trained and simulated, with subsequent short-term incidence rates (Jan-Jun 2003) forecasted and compared to actual data.
• The RBF network demonstrated superior effectiveness and feasibility for short-term infectious disease incidence prediction compared to the ARIMA model.

Impact:

• The study highlights the potential of RBF neural networks as a powerful tool for infectious disease surveillance and public health planning.
• Findings suggest that RBF models can improve the accuracy of short-term infectious disease forecasting, aiding in resource allocation and intervention strategies.
• This research contributes to the integration of advanced computational methods in epidemiology for better disease management.