Forecasting Trends in the Number of Tuberculosis Cases in China Using the SARIMA Model

Development of a seasonal ARIMA model to forecast the trend and number of tuberculosis cases in China. It is useful for analysing tuberculosis prevention and control, treatment and the detection of related external variables from public health dimension. SARIMA model were used to modelling data on tuberculosis incidence cases in China from 2004 to 2018, and comparing and analysing the fitted values of the model with the actual values. The ARIMA (4, 1, 0) (0, 1, 1) [12] function has a MAPE of only 4.07 and a value of 0.89 for the R-squared, which is able to fit the 2004-2018 TB case data well. The error rate in January, the peak month of incidence, was 2.81%. The error rate in October, the turning point of the incidence rate, was 0.08 %. As a predictive model, this is a relatively good fit goodness of fit reference value. This result can provide time-scale suggestions for the development of TB prevention and control measures, the arrangement of medical resources, and the production of relevant drugs and medical devices.