Homogeneity detection for the high-dimensional generalized linear model

被引：9

作者：

Jeon, Jong-June ^{[1
]}

Kwon, Sunghoon ^{[2
]}

Choi, Hosik ^{[3
]}

机构：

[1] Univ Seoul, Dept Stat, Seoul 02504, South Korea

[2] Konkuk Univ, Dept Appl Stat, Seoul 05029, South Korea

[3] Kyonggi Univ, Dept Appl Stat, Suwon 16227, South Korea

来源：

COMPUTATIONAL STATISTICS & DATA ANALYSIS | 2017年 / 114卷

基金：

新加坡国家研究基金会;

关键词：

Categorical covariates; Generalized linear model; Grouping penalty; Oracle property; NONCONCAVE PENALIZED LIKELIHOOD; VARIABLE SELECTION; GROUPING PURSUIT; REGRESSION; SHRINKAGE; LASSO;

D O I：

10.1016/j.csda.2017.04.001

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

We propose to use a penalized estimator for detecting homogeneity of the high dimensional generalized linear model. Here, the homogeneity is a specific model structure where regression coefficients are grouped having exactly the same value in each group. The proposed estimator achieves weak oracle property under mild regularity conditions and is invariant to the choice of reference levels when there are categorical covariates in the model. An efficient algorithm is also provided. Various numerical studies confirm that the proposed penalized estimator gives better performance than other conventional variable selection estimators when the model has homogeneity. (C) 2017 Elsevier B.V. All rights reserved.

引用

页码：61 / 74

页数：14

共 46 条

[1] The DC (difference of convex functions) programming and DCA revisited with DC models of real world nonconvex optimization problems [J].