摘要
目的应用灰色模型[GM（1，1）]和ARIMA乘积季节模型，预测承德市布鲁菌病的发病趋势，比较两种模型预测的效果。方法利用承德市2008-2014年布鲁菌病的统计结果，建立GM（1，1）模型和ARIMA乘积季节模型，分别预测2015年1-12月的发病数，并将结果与实际监测结果进行比较，用平均相对误差验证模型的可靠性。结果GM（1，1）模型为X^0（k）=0．0016（X^0＋23712．31）exp[0．0016（k-t^0）]，ARIMA乘积季节的最优模型为ARIMA（0，1，1）×（0，1，0）。两种模型预测的2015年1-12月发病数与实际监测数比较得到平均相对误差分别为114．82％、11．66％。结论ARIMA乘积季节模型对布鲁菌病的预测效果较好，可以用于短期预测，预测结果可作为科学防控的理论依据。
Objective The grey model [GM （1,1）] and the ARIMA multiple seasonal model were used to predict the incidence trend of brucellosis in Chengde, and the effects of predictions of the two models is compared. Methods According to the statistical results of epidemiological research results of monthly brucellosis patients in Chengde from 2008 to 2014, we established the ARIMA multiple seasonal model and GM （1,1） model, individually predicted the trend of brucellosis in 2015. Compared with the actual monitoring results, the average relative error was used to verify the reliability of the model. Results GM （1,1） model was X^0（k） = 0.001 6 （X^0 ＋ 23 712.31） exp [0.001 6 （k - t0）], and the optimal model for determining the ARIMA multiple seasonal model of product was ARIMA （0, 1, 1） × （0, 1, 0）12. The average relative errors of the two models predicted in 2015 from January to December compared with the actual monitoring data were 114.82% and 11.66%. Conclusion The product ARIMA multiple seasonal model has better predictive effect on brucellosis, and can be used for short-term forecasting.
作者
何海阔
王占辉
He Haikuo;Wang Zhanhui(Department of Computer Science, Chengde Petroleum College, Chengde 067000, China （He HK;Physics and Chemistry Department, Chengde Center for Disease Control and Prevention, Chengde 067000, China （Wang ZH)
出处
《中华地方病学杂志》
CAS
CSCD
北大核心
2018年第4期338-340,共3页
Chinese Journal of Endemiology
基金
承德市科技支撑计划项目（201601A001、201701A005）
