The spatio-temporal analysis and prediction model comparison of incidence rate of other infectious diarrhea diseases in Qinghai Province from 2017 to 2023
Objective To analyze the epidemiological trends and characteristics of other infec-tious diarrheal diseases(OIDD)in Qinghai Province,and to provide predictions for these diseases in Qinghai Province for 2024.Methods Using monthly and annual incidence rates of OIDD in Qinghai Province from January 2017 to December 2023 as primary data,the study employed ArcGIS 10.8 software for map visualization of annual incidence rates in Qinghai Province,and GeoDa 1.16 software for spatial autocorrelation analysis.R 4.3.1 software was used to construct various models for OIDD in Qinghai Prov-ince,including seasonal autoregressive integrated moving average(SARIMA)model,triple exponential smoothing(Holt-Winters)model,neural network autoregression(NNAR)model,trigonometric seasonal-ity,Box-Cox transformation(TBATS)model,and Prophet model.The models'fitting effects were evalua-ted using root mean square error(RMSE),mean absolute error(MAE),and mean absolute percentage error(MAPE).Results All models,except the Holt-Winters model,effectively captured the incidence rate trends.Among them,the NNAR model performed best in the training set,with MAE of 0.90,RMSE of 1.25,and MAPE of 16.43,outperforming models such as TBATS.In the test set,while its RMSE val-ue was higher than those of the SARIMA and TBATS models,its MAE and MAPE values were lower than other models,indicating the best overall predictive performance.Therefore,the NNAR model can be used to forecast the incidence rate of OIDD in Qinghai Province for 2024,providing insights for disease preven-tion strategies in high-altitude regions.Conclusions From 2017 to 2023,Xining City,Haidong City,and Huangnan Tibetan Autonomous Prefecture in Qinghai Province were high-incidence areas for OIDD.Among the predictive models,the NNAR model showed the best performance.However,in practical ap-plications,it is necessary to develop corresponding prevention and control measures by considering the spatiotemporal characteristics and epidemic trends of each region.
Other infectious diarrhea diseaseNeural network autoregression modelModel pre-dictionSeasonal autoregressive integrated moving average modelProphet model
万方数据
