基于小波分析的肾综合征出血热发病率预测方法  

The Method of Wavelet-based Analysis and Forecasting for Haemorrhagic Fever with Renal Syndrome

作　　者：吴学森[1] 王洁贞[1] 刘云霞[1] 张娜[1] 

机构地区：[1]山东大学公共卫生学院卫生统计教研室

出　　处：《中国卫生统计》2005年第1期9-12,共4页Chinese Journal of Health Statistics

基　　金：国家自然科学基金资助 (30 0 70 677)

摘　　要：目的　建立季节性趋势时间序列小波预测模型的方法 ,提高肾综合征出血热发病率的预测精度。方法　对原始序列进行多层小波分解 ,提取趋势项、周期项和随机项 ,对它们分别进行预测 ,将各项的预测值合并作为原序列的预测值。结果　小波分解后建模的预测精度为 84.8% ,而ARIMA建模的预测精度为 66.1%。结论　用小波预测模型对HFRS作短、中期预测是有效。Objective To improve the forecasting precision of the incidence rate for Haemorrhagic Fever with Renal Syndrome(HFRS) ,this paper proposed the forecasting method of the seasonal tendency time series called wavelet forecasting model.Methods By wavelet decomposing,the trend term,periodical term and stochastic term of the original series were separated and modelinged respectively.The final forecasting results are composed of these three terms'forecasting values.Results The forecasting precision of wavelet forecasting model and ARIMA model is 84.8% and 66.1% respectively.Conclusion Wavelet forecasting model is effective and feasible for HFRS's incidence rate prediction in the short and the middle term.

关 键 词：肾综合征出血热 发病率 HFRS 序列 随机 季节性 方法 预测精度 ARIMA 合并 

分 类 号：R181.3[医药卫生—流行病学]