一种径流随机模拟的非参数模型  

A Kind of Nonparametric Model for Runoff Stochastic Generation

作　　者：王文圣[1] 马吉让 向红莲[1] 丁晶[1] 

机构地区：[1]四川大学水利水电工程学院,四川成都610065 [2]黄河山东水资源保护局,山东济南250100

出　　处：《水利水电技术》2002年第2期8-10,共3页Water Resources and Hydropower Engineering

基　　金：国家自然科学基金重大项目(50099620);四川大学高速水力学国家重点实验室开放基金(2008);四川大学青年科学研究基金(2001)资助.

摘　　要：应用核密度估计理论构造了单变量多阶非参数模型NP(p).该模型是基于数据驱动的、不需对序列概率分布和相依形式进行识别、适合于时间序列随机模拟的非参数模型.将NP(p)模型用于金沙江流域李庄-屏山区间年径流过程随机模拟,与平稳AR(p)模型对比,研究表明非参数模型是适合于径流随机模拟的.In this paper,a kind of nonparametric model based on the kernel density estimation theories is constructed for time series of single variable.It belongs to a class of data-driven approach and avoids the form of probability distribution and the form of dependence.The model has clear concept and single structure.The model was applied to the stochastic generation of average runoff time series at the region of Li zhuang-Pingshan.The results from model verification and validation have shown that the suggested model is superiority to AR(p)model and functional for stochastic simulation of hydrology time series.

关 键 词：核密度估计 非参数模型 自回归模型 随机模拟 径流 单变量多阶 

分 类 号：P333[天文地球—水文科学]