摘要
目的探讨3种不同的模型在肾综合征出血热(HFRS)发病率拟合预测中的应用,并选用合适的模型预测HFRS在该地区未来的发病趋势,为合理调配HFRS防治的卫生资源提供科学依据。方法采用灰色GM(1,1)模型、自回归模型、ARIMA模型对1990~2007年沈阳市HFRS的发病率资料进行数据拟合,并比较3个模型的拟合效果,选择最优模型预测沈阳地区未来几年的HFRS发病趋势。结果针对沈阳市HFRS发病率建立的GM(1,1)模型、自回归模型和ARIMA模型的平均误差率(MER)分别为52.76%、20.53%和6.75%,R2分别为0.466、0.945和0.991;预测在2012年前后沈阳市HFRS发病将会出现一个高峰,达到4.4035/10万。结论对于隐含波动周期并且不稳定的循环型时间序列,无论拟合还是预测,ARIMA模型的效果都优于灰色GM(1,1)模型和自回归模型。目前沈阳市HFRS正处于发病率的低谷期,预测未来几年呈上升趋势,应引起注意。
Objective To explore the applications of three different models in prediction of incidence rate of Hemorrhagic Fever with Renal Syndrome(HFRS) ,and to predict the incidence trend of the disease in later years in Shenyang so as to provide scientific basis for reasonable distribution of health resource of HFRS prevention and treatment. Methods The data of the incidence rate in Shenyang from 1990 to 2007 were fitted using GM ( 1,1 ) Grey model,Auto-regression model and ARIMA model, the effects of tilting were compared, and the incidence rate of HFRS in later years was predicted. Results The MER of GM ( 1,1 ) Grey .model, Auto-regression model and ARIMA model in Shenyang was 52.76%, 20.53% and 6.75%, respectively, and the R2 of the three models was 0.466,0.945 and 0.991, respectively. The incidence rate of HFRS of Shenyang would again reach a new peak of 4.4035 per 10 million in 2012. Conclusion Compared with the GM ( 1,1 ) Grey model and Auto-regression model, ARIMA model is a good choice in unstable and recycling time series data on fitting and forecasting. At present,the morbidity of HFRS is in low tide in Shenyang city,but it shows the trend of rising.We must pay more attention and take some effective methods to prevent its incidence.
出处
《中国医科大学学报》
CAS
CSCD
北大核心
2008年第6期843-846,共4页
Journal of China Medical University
基金
国家自然科学基金资助项目(30771860)
(70503028)