基于小波消噪的改进的非参数解集模型  

Modified Nonparametric Disaggregation Model Based on Wavelet De-noise

作　　者：赵太想[1] 王文圣[1] 丁晶[1] 

机构地区：[1]四川大学水电学院,四川成都610065

出　　处：《四川大学学报（工程科学版）》2005年第5期1-4,共4页Journal of Sichuan University (Engineering Science Edition)

基　　金：国家自然科学基金资助项目(50279023);四川省科技厅软科学基金资助项目(042R025-051)

摘　　要：首先对文献[2]提出的非参数解集模型进行了改进,然后建议了用小波技术对水文序列进行消噪处理再估计非参数解集模型中的带宽系数的策略。改进的非参数解集模型无需对水文序列相依结构(线性或非线性)和概率分布形式(正态分布、P-III型分布)作假定,且能保持水文序列的各种统计特性,适用于时间量和空间量的随机解集。最后以金沙江流域某站月径流随机模拟为例进行验证,研究表明:改进模型能反映实测月径流系列的变化特性,因而用于水文随机模拟是可行的。Firstly, based on the nonparametric disaggregation model in reference[2], a modified nonparametric disaggregation model is presented. Then a estimation of the band-width coefficient h of nonparametric disaggregafion model is given with wavelet de-noise technology. The modified nonparametric disaggregation model has no prior assumptions as to the form of dependence and the form of probability density function, and can preserve all statistical properties of real hydrological time series. At the same time the suggested model is suitable to the hydrological variables in time and space. Lastly, The results of monthly nmoff stochastic simulation at a hydrological station in Jing Sha River show that the modified model works well.

关 键 词：非参数解集模型 小波消噪 核密度估计 随机模拟 

分 类 号：P333[天文地球—水文科学]