带隐变量的回归模型EM算法

计算机科学 ›› 2014, Vol. 41 ›› Issue (2): 136-140.

带隐变量的回归模型EM算法

韩忠明,吕涛,张慧,姜同强

北京工商大学计算机与信息工程学院北京100048;北京工商大学计算机与信息工程学院北京100048;北京工商大学计算机与信息工程学院北京100048;北京工商大学计算机与信息工程学院北京100048

出版日期:2018-11-14 发布日期:2018-11-14
基金资助:
本文受国家自然科学基金(61170112),北京市属高等学校科学技术与研究生教育创新工程建设项目(PXM2012_014213_000037)资助

EM Algorithm for Latent Regression Model

HAN Zhong-ming,LV Tao,ZHANG Hui and JIANG Tong-qiang

Online:2018-11-14 Published:2018-11-14

摘要/Abstract

摘要： 带有隐变量的回归模型具有非常广泛的应用场合,隐回归模型的参数求解问题依赖于自变量的分布假设。基于自变量的beta分布的假设条件,给出了隐回归模型的EM算法,详细地推导了模型中的参数求解过程,给出了使用牛顿法求解beta分布参数的算法,并提出一个合适的初值选择算法。在模拟数据和真实数据的基础上进行了详细的比较性试验,结果表明,对具有不同分布特征的因变量观察值,EM算法能够有效地求解隐回归模型的参数。

关键词: 隐回归模型,最大期望算法,回归模型中图法分类号TP391．4文献标识码A

Abstract: There have a very wide range of applications for latent variable regression model．Estimation of the parameters of latent variable regression models depends on the assumptions of the distribution of the independent variables．Based on the Beta distribution of the independent variables,an EM algorithm for parameters estimation of latent regression model was proposed in this paper．The detailed solution process in the model was derived．Newton method for solving parameter of Beta distribution was given．Furthermore,an initial value selection algorithm was proposed．Comprehensive experiments were conducted based on simulation datasets and real dataset．The experimental results show that the EM algorithm can efficiently estimate parameters with different distribution shapes of latent regression models.

Key words: Latent regression model,EM algorithm,Regression model

韩忠明,吕涛,张慧,姜同强. 带隐变量的回归模型EM算法[J]. 计算机科学, 2014, 41(2): 136-140. https://doi.org/

HAN Zhong-ming,LV Tao,ZHANG Hui and JIANG Tong-qiang. EM Algorithm for Latent Regression Model[J]. Computer Science, 2014, 41(2): 136-140. https://doi.org/

参考文献

[1] Tarpey T,Petkova E.Latent regression analysis[J].Statistical Modelling,2010,0(2):133-158
[2] Moustaki I,Knott M.Generalized latent trait models[J]．Psychometrika,2000,65(3):391-411
[3] Aitchison J,Shen S M．Logistic-normal distributions:Some pro-perties and uses[J].Biometrika,1980,67(2):261-72
[4] Barto lucci F,Scaccia L．The use of mixtures for dealing withnon-normal regression errors [J]．Computational Statistics & data Analysis,2005,48(4):821-48
[5] Dorazio R M,Andrew Royle J．Mixture models for estimating the size of a closedpopulation when capture rates vary among individuals[J]．Biometrics,2003,59(2):351-64
[6] Khan M E,Mohamed S,Marlin B M,et al.A stick-BreakingLikelihood for categorical data analysis with latent Gaussian models[C]∥Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS).2012:610-618
[7] Irncheeva I,Cantoni E,Genton M G．Generalized linear latent variable models with flexible distribution of latent variables[J]．Scandinavian Journal of Statistics,2012,39(4):663-680
[8] Kelava A,Kohler M,Krzyzak A,et al．Nonparametric estimation of a latent variable model．http://www3.mathematik.tu-darmstadt.de/hp/stochastik-homepages/kohler-michael/publika-tionen.html,2012
[9] Guo J,Wall M,Amemiya Y．Latent class regression on latent factors[J].Biostatistics,2006,7(1):145-63
[10] Tarpey T,Ivey C T．Allometricex tension for multivariate regression models[J]．Journal of Data Science,2006,4(4):479-95
[11] Bartoletti S,Flury B D,Nel D G．Allometric extension[J].Biometrics,1999,5(4):1210-1214
[12] Tarpey T,Yun D,Petkova E．Model misspecification finite mixture or homogeneous?[J].Statistical modeling,2008,8(2):199-218
[13] Tarpey T,Petkova E．Modeling Placebo Response via InfiniteMixtures[J]．Jpn Journal of Biostatistics,2010,4(2):161-179
[14] Tarpey T,Petkova E,Lu Y,et al．Optimal partitioning for linear mixed effects models:Applications to identifying placebo responders[J].Journal of the American Statistical Association,2010,5(491):968-977
[15] Tsou C M．On the exploration of linear latent effect for multivariate modeling[J].Applied Mathematical Modelling,2012,6(12):6154-6166
[16] Ma Y,Genton M G．Explicit estimating equations for semiparametric generalized linear latent variable models[J]．Journal of the Royal Statistical Society:Series B (Statistical Methodology),2010,2(4):475-495
[17] 韩忠明,苑丽玲,杨伟杰,等．加权社会网络中重要节点发现算法[J]．计算机应用,2013,3(6):1553-1557

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed