非参数模型的三种经典估计方法  

Three Powerful Estimations for Non-parametric Model

作　　者：朱桂玲[1] 李治远 

机构地区：[1]云南大学数学与统计学院,云南昆明650500

出　　处：《保山学院学报》2016年第5期57-59,共3页JOURNAL OF BAOSHAN UNIVERSITY

基　　金：国家自然科学基金项目(项目编号:11561075)

摘　　要：参数估计是在概率密度函数已知的前提下,利用似然函数对参数进行估计,但是,在实际问题中,很多数据的分布是事先无法假定的,于是非参数模型就应运而生;主要利用核估计、局部多项式估计和k近邻估计三种经典方法对非参数模型进行估计,并辅以经典的例题;最后,通过一个综合模拟计算对这三种估计方法进行了比较并证明所提出的方法是有效和可行的。Parameter estimation is to estimate parameter with likelihood function when the probability density function is known. However, in practical problems, we can＇t assume many distributions of the data in advance, so the non-parametric model is made. In this paper, we mainly use three methods to estimate non-parametric model, which are Kernel estimation, local polynomial estimation, k nearest neighbor estimation. Behind each of the estimation method followed a typical example. Finally, a simulation was compared these three estimation methods and have shown that the proposed methods are effective and feasible.

关 键 词：核估计 局部多项式估计 k近邻估计 非参数模型 

分 类 号：O13[理学—数学]