Abstract：Objective To establish a time series model suitable for predicting the number of cases of brucellosis in China, and provide a scientific basis for preventing and controlling this disease.Methods This study utilized data on the number of cases from January 2015 to December 2021 to establish SARIMA and prophet models, respectively. Subsequently, the predictive performance of these two models was validated by using data from January 2022 to April 2023, and the results of the model predictions were compared by using three indicators: RMSE, MAPE, and MAE. Finally, the prophet model with higher accuracy was used to predict the number of cases from May 2023 to April 2024.Results The overall number of cases of brucellosis in China is rising, reaching its peak from June to July each year, showing a clear seasonal trend. Compared to the SARIMA model, the RMSE, MAPE, and MAE values of the prophet model are lower, indicating that the prophet model has higher accuracy in predicting brucellosis. The predicted peak of the number of cases in 2023 is lower than the actual peak in 2021 and 2022.Conclusion The prophet model can better fit the monthly reported number of cases of brucellosis in the country and can be used for short-term prediction.
温福东，赵彬宇，苏月，王玉鹏. 基于prophet模型预测中国布鲁氏菌病发病人数[J]. 中国医院统计, 2024, 31(1): 7-10.
Wen Fudong，Zhao Binyu，Su Yue，Wang Yupeng. Prediction of the number of brucellosis cases in China based on prophet model. journal1, 2024, 31(1): 7-10.
［2］ZEINALI M, DOOSTI S, AMIRI B, et al. Trends in the epidemiology of brucellosis cases in Iran during the last decade［J］. Iran J Public Health, 2022, 51(12):2791-2798. DOI: 10.18502/ijph.v51i12.11470.
