A Gaussian Process Regression and Wavelet Transform Time Series approaches to modeling Influenza A

The global spread of Influenza A viruses is worsening economic and social challenges. Various mechanistic models have been developed to understand the virus's spread and evaluate intervention effectiveness. This study aimed to model the temporal dynamics of Influenza A using Gaussian Process Regression (GPR) and wavelet transform approaches. The study employed Continuous Wavelet Transform (CWT), Discrete Wavelet Transform (DWT) and Wavelet Power Spectrum to analyze time-series data from 2009 to 2023. The GPR model, known for its non-parametric Bayesian nature, effectively captured non-linear trends in the Influenza A data, while wavelet transforms provided insights into frequency and time-localized characteristics. The integration of GPR with DWT denoising techniques demonstrated superior performance in forecasting Influenza A cases compared to traditional models like Auto Regressive Integrated Moving Averages (ARIMA) and Exponential Smoothing (ETS) using Holt-Winter method. The study identified significant anomalies in Influenza A cases, corresponding to known pandemic events and seasonal variations. These findings highlight the effectiveness of combining wavelet transform analysis with GPR in understanding and predicting infectious disease patterns, offering valuable insights for public health planning and intervention strategies. The research recommends extending this approach to other respiratory viruses to assess its broader applicability.