|本期目录/Table of Contents|

[1]张煜东,颜俊,王水花,等.非参数估计方法[J].武汉工程大学学报,2010,(07):99-106.[doi:10.3969/j.issn.16742869.2010.07.025]
　ZHANG Yu dong,YAN Jun,WANG Shui hua,et al.Survey of nonparametric estimation methods[J].Journal of Wuhan Institute of Technology,2010,(07):99-106.[doi:10.3969/j.issn.16742869.2010.07.025]
点击复制


非参数估计方法(/HTML)
分享到：

《武汉工程大学学报》[ISSN:1674-2869/CN:42-1779/TQ]

卷:
期数:
2010年07期
页码:
99-106
栏目:
机电与信息工程
出版日期:
2010-07-31

文章信息/Info

Title:
Survey of nonparametric estimation methods
文章编号:
16742869(2010)07009908
作者:
张煜东1; 2; 颜俊1; 王水花1; 吴乐南1
1.东南大学信息科学与工程学院，江苏 南京 210096;
2.哥仑比亚大学精神病学系脑成像实验室,纽约州 纽约 10032
Author(s):
ZHANG Yudong1; 2; YAN Jun1; WANG Shuihua1; WU Lenan1
1.School of Information Science & Engineering, Southeast University, Nanjing 210096,China;
2.Brainimaging Lab., Department of Psychology, Columbia University, New York NY 10032,USA
关键词:
参数统计; 非参数统计; 核方法; 局部多项式回归; 正则化方法; 正态均值模型; 小波; 超完备字典; 前向神经网络; 径向基函数网络
Keywords:
 ; parametric statistics; nonparametric statistics; kernel method; local polynomial regression; regularization method; normal mean model; wavelet; overcomplete dictionary; forward neural network; radial basis function network
分类号:
O212.7
DOI:
10.3969/j.issn.16742869.2010.07.025
文献标志码:
A
摘要:
为了解决函数估计问题，首先讨论了传统的参数回归方法.由于传统方法需要先验知识来决定参数模型，因此不稳健，且对模型敏感.因此，引入了基于数据驱动的非参数方法，无需任何先验知识即可对未知函数进行估计.本文主要介绍最新的8种非参数回归方法：核方法、局部多项式回归、正则化方法、正态均值模型、小波方法、超完备字典、前向神经网络、径向基函数网络.比较了不同的算法，给出算法之间的相关性与继承性.最后，将算法推广到高维情况，指出面临计算的维数诅咒与样本的维数诅咒两个问题.通过研究指出前者可以通过智能优化算法求解，而后者是问题固有的.
Abstract:
In order to solve the problem of function estimation, we first discuss traditional parametric regression method.Since it needs a priori knowledge to determine the model, the parametric method is not robust and is modelsensitive.Thus, datadriven nonparametric method is introduced, which needs not any a prior knowledge to estimate the unknown function.Eight major nonparametric methods are discussed as kernel method, local polynomial regression, regularization method, normal mean model, wavelet method, overcomplete dictionary, forward neural network, and radial basis function network.These algorithms are compared, and their coherence and inheritance are investigated.Finally, generalize the algorithms to high dimensionality and point out two problems as curse of dimensionality of computation and sample.The former can be settled down by intelligent methods while the latter is problem intrinsic.

参考文献/References:

［1］Neumeyer N.A note on uniform consistency of monotone function estimators ［J］. Statistics & Probability Letters,2007,77(7):693703
［2］Sheena Y,Gupta A K.New estimator for functions of the canonical correlation coefficients ［J］. Journal of Statistical Planning and Inference,2005,131(1):4161.
［3］张煜东,吴乐南,李铜川,等.基于PCNN的彩色图像直方图均衡化增强［J］.东南大学学报，2010,40(1):6468.
［4］詹锦华.基于优化灰色模型的农村居民消费结构预测［J］.武汉工程大学学报，2009,31（9）:8991.
［5］Wasserman L. All of Nonparametric Statistics ［M］.New York:SpringerVerlag, Inc.
［6］张煜东, 吴乐南, 吴含前.工程优化问题中神经网络与进化算法的比较［J］.计算机工程与应用,2009,45(3):16.
［7］Hansen C B.Asymptotic properties of a robust variance matrix estimator for panel data when T is large ［J］.Journal of Econometrics,2007,141(2):597620.
［8］Pokharel P P, Liu W F, Principe J C.Kernel least mean square algorithm with constrained growth ［J］.Signal Processing,2009,89(3):257265.
［9］Kalivas J H.Cyclic subspace regression with analysis of the hat matrix ［J］.Chemometrics and Intelligent Laboratory Systems,1999,45(1):215224.
［10］张煜东,吴乐南.基于二维Tsallis熵的改进PCNN图像分割［J］.东南大学学报：自然科学版,2008,38(4):579584
［11］Gekinli N C, Yavuz D.A set of optimal discrete linear smoothers［J］.Signal Processing,2001,3(1):4962.
［12］Antoniotti M,Carreras M,Farinaccio A,et al.An application of kernel methods to gene cluster temporal metaanalysis ［J］.Computers & Operations Research,2010,37(8):13611368.
［13］Hsieh P F,Chou P W,Chuang H Y.An MRFbased kernel method for nonlinear feature extraction ［J］.Image and Vision Computing,2010,28(3):502517.
［14］Katkovnik V.Multiresolution local polynomial regression:A new approach to pointwise spatial adaptation ［J］.Digital Signal Processing,2005,15(1):73116.
［15］Baíllo A,Grané A.Local linear regression for functional predictor and scalar response ［J］.Journal of Multivariate Analysis,2009,100(1):102111.
［16］Zhang J W,Krause F L.Extending cubic uniform Bsplines by unified trigonometric and hyperbolic basis ［J］.Graphical Models,2005,67(2):100119.
［17］张煜东,吴乐南,韦耿,等.用于多指数拟合的一种混沌免疫粒子群优化［J］.东南大学学报,2009,39(4):678683.
［18］Chaudhuri S,Perlman M D.Consistent estimation of the minimum normal mean under the treeorder restriction ［J］.Journal of Statistical Planning and Inference,2007,137(11):33173335.
［19］Labat D.Recent advances in wavelet analyses:Part 1.A review of concepts［J］.Journal of Hydrology,2005,314(1):275288.
［20］Kunoth A.Adaptive Wavelets for Sparse Representations of Scattered Data［J］.Studies in Computational Mathematics,2006,12:85108.
［21］Donoho D L, Elad M.On the stability of the basis pursuit in the presence of noise［J］.Signal Processing,2006,86(3):511532.
［22］Malgouyres F.Rank related properties for Basis Pursuit and total variation regularization ［J］.Signal Processing,2007,87(11):26952707.
［23］张煜东,吴乐南,韦耿.神经网络泛化增强技术研究［J］.科学技术与工程,2009,9(17):49975002.
［24］屠艳平,管昌生,谭浩.基于BP网络的钢筋混凝土结构时变可靠度［J］.武汉工程大学学报,2008,30（3）:3639.
［25］Zhang Y D,Wu L N,Neggaz N, et al.Remotesensing Image Classification Based on an Improved Probabilistic Neural Network［J］.Sensors,2009,9:75167539.
［26］Aleksandrowicz G,Barequet G.Counting polycubes without the dimensionality curse ［J］.Discrete Mathematics,2009,309(13):45764583.
［27］张煜东,吴乐南,奚吉,等.进化计算研究现状（上）［J］.电脑开发与应用,2009,22(12):15.
［28］王忠,叶雄飞.遗传算法在数字水印技术中的应用［J］.武汉工程大学学报,2008,30(1):9597.

相似文献/References:

备注/Memo

备注/Memo:
收稿日期：20100402基金项目：国家自然科学基金（60872075）；国家高技术发展计划（2008AA01Z227）；高等学校科技创新工程重大项目培育资金项目（706028）作者简介：张煜东(1985)，男，江苏苏州人，哥伦比亚大学博士后.研究方向：人工智能、数据挖掘、脑图像处理.
更新日期/Last Update: