Influenza incidence trend prediction based on ARIMA seasonal multiplicative model

Abstract

This study utilizes a time series ARIMA seasonal multiplicative model to predict the incidence trend of influenza in China, providing valuable early warning references for influenza prevention and control. Monthly incidence data of influenza cases nationwide were collected from January 2014 to August 2022. The data from January 2014 to December 2020 were used as the training set to fit the time series model of influenza incidence. The data from January 2021 to August 2022 were utilized as the testing set to predict the influenza incidence from January 2021 to August 2022 using the fitted model. The predicted values were then compared with the testing set. Through residual white noise testing, significance testing of parameters, and examination of model fit, the final model was determined as , which demonstrated a good fit. The majority of actual data fell within the 95% confidence interval of the predicted values, and the predicted incidence trend aligned closely with the actual trend. The constructed model holds significant application value in early warning systems for influenza prevention and control, providing crucial insights for public health strategies. In practical applications, this model can be integrated with various factors such as social, natural, and geographic environments to formulate targeted prevention and control strategies, thus enhancing the efficiency and effectiveness of influenza prevention and control measures.