应用GM(1,1,sinω)模型预测肾综合征出血热发病率研究  

Forecasting incidence of hemorrhagic fever with renal syndrome using the GM (1,1,sinω) model

作　　者：郭立春[1] 王汉宁[2] 吴伟[1] 郭军巧[3] 王萍[4] 周宝森[1] 

机构地区：[1]中国医科大学流行病学教研室,辽宁沈阳110001 [2]中国医科大学,91期辽宁沈阳110001 [3]辽宁省疾病预防控制中心,辽宁沈阳110001 [4]沈阳市疾病预防控制中心,辽宁沈阳110001

出　　处：《中国热带医学》2009年第3期415-417,共3页China Tropical Medicine

基　　金：国家自然科学基金资助项目(30771860)

摘　　要：目的探讨GM(1,1,sinω)模型在肾综合征出血热(HFRS)发病率预测的应用。方法利用1984～2004年沈阳市HFRS发病率资料建立GM(1,1)预测模型和GM(1,1,sinω)预测模型,对样本进行拟合和预测并对两者的拟合和预测效果进行比较。结果GM(1,1)预测模型为XΣ(1)(k+1)=-541.5277e-0.0092k+551.4778;GM(1,1,sinω)模型为XΣω(1)(k+1)=-158.4104e-0.0444k+162.6622+11.7276sin2kπ/21+5.6982cos2kπ/21,GM(1,1,sinω)模型拟合精度较好(C=0.3912,P=0.9048)。GM(1,1)和GM(1,1,sinω)预测模型拟合的平均误差率(MER)分别为50.22%、20.34%;两者的预测MER分别为25.64%、13.10%,无论从拟合效果还是从预测效果来看GM(1,1,sinω)模型xing1,sinω)forecast的MER均低于GM(1,1)模型。结论GM(1,1,sinω)模型克服了传统灰色模型GM(1,1)的局限性,对于波动性较大且具有周期性的资料具有很好的实用价值。Objective To discuss the application of GM （1,1,sinω） model in forecasting the incidence of hemorrhagic fever with renal syndrome （HFRS）. Methods GM （1,1） forecasting model and GM （1,1,sinω） forecasting model were established by using incidence data of HFRS in Shenyang City from 1984 to 2004 . The damples were Fitted and forecasted （1） and the results were compared. Results The GM（1,1） forecasting model was X^（1） （k＋1） =-541.5277e^-0.0092k＋551.4778; the （1） GM （1,1,sinω） forecasting model wasXω^（1） （k＋1） =-158.4104e^-0.0444k＋162.6622＋11.7276 sin2kπ/21＋5.6982cos2kπ/21.The GM （1,1,simω） model was fitted better （C=0.3912,P=0.9048）. The MERs of fitness of GM（1,1） model and GM（1,1,sinω） model were 50.22% and 20.34% respectively; the MERs of forecasting of them were 25.64% and 13.10%, respectively . The MER of GM （1,1,sinω） model was less than that of GM （1,1） model, either in fitting effect or the predictive effect. Conclusion The GM （1,1,sinω） model overcomes the limitations of the traditional grey model GM （1,1）, and has a good applied value, for the fluctuated and periodical data.

关 键 词：灰色模型 肾综合征出血热 预测 

分 类 号：R512.8[医药卫生—内科学]