Optimized Forecasting Method for Weekly Influenza Confirmed Cases.

Area of Science:

• Epidemiology and Public Health
• Computational Intelligence and Machine Learning

Background:

• Influenza epidemics pose a significant global threat, necessitating accurate forecasting for effective public health interventions.
• Traditional forecasting models often face limitations in capturing complex temporal dynamics of infectious disease outbreaks.

Purpose of the Study:

• To develop and evaluate an enhanced Adaptive Neuro-Fuzzy Inference System (ANFIS) for forecasting weekly confirmed influenza cases.
• To improve the predictive performance of ANFIS by integrating metaheuristic optimization algorithms, specifically Flower Pollination Algorithm (FPA) and Sine Cosine Algorithm (SCA).

Main Methods:

• An enhanced ANFIS model, termed FPASCA-ANFIS, was developed by optimizing ANFIS parameters using the FPA and SCA metaheuristics.
• The model was trained and validated using official weekly influenza case data from China and the USA, sourced from CDC and WHO.
• Performance was benchmarked against existing state-of-the-art forecasting approaches using metrics like RMSRE, MAPE, MAE, and R².

Main Results:

• The proposed FPASCA-ANFIS model demonstrated superior performance in forecasting weekly influenza cases compared to conventional methods.
• Quantitative evaluation using multiple error metrics (RMSRE, MAPE, MAE) and accuracy measure (R²) confirmed the enhanced model's effectiveness.
• The integration of FPA and SCA significantly improved the ANFIS model's ability to predict influenza incidence.

Conclusions:

• The FPASCA-ANFIS model offers a robust and accurate approach for forecasting influenza epidemics.
• This enhanced computational intelligence technique can aid public health organizations in developing timely and effective disease control strategies.
• Accurate influenza forecasting is vital for resource allocation and policy-making during public health emergencies.