Prediction of Influenza in Henan Province Based on SARIMAX-LSTM Hybrid Model

In this paper, the relevant data of influenza-like case rate, resident consumer index, number of vaccinated and weather in Henan province in the past decade were selected, and the missing value and Pearson correlation analysis were processed through Lagrange interpolation to ensure the integrity and stability of the data. In this paper, first using the SARIMA model to analyze the data, Can effectively capture the seasonal variation in the data, However, it shows some limitations when dealing with nonlinear characteristics and external influence factors; Secondly, the SARIMAX model was introduced, Adding the significant variable temperature as an external factor, However, there is still room for optimization for complex nonlinear dynamics; Finally, by further introducing the LSTM model, Form a SARIMAXLSTM hybrid model, The empirical results show that, When the mixed model predicts the changing trend in the data, Showed higher accuracy and adaptability than the single model and decreased seasonal influenza activity during the summer, Gradually increasing in autumn and winter. The research results provide a scientific basis for public health prevention and control, and reflect the importance of data analysis in the social process.