Variable Selection of Generalized Regression Models Based on Maximum Rank Correlation  

作　　者：Peng-jie DAI Qing-zhao ZHANG Zhi-hua SUN 

机构地区：[1]School of Business,Renmin University of China [2]Department of Mathematics,University of Chinese Academy of Sciences

出　　处：《Acta Mathematicae Applicatae Sinica》2014年第3期833-844,共12页应用数学学报（英文版）

基　　金：supported by National Natural Science Foundation of China(10901162);supported by the Fundamental Research Funds for the Central Universities and the Research Funds of Renmin University of China(10XNF073);supported by China Postdoctoral Science Foundation(2014M550799)

摘　　要：In this paper, we investigate the variable selection problem of the generalized regression models. To estimate the regression parameter, a procedure combining the rank correlation method and the adaptive lasso technique is developed, which is proved to have oracle properties. A modified IMO （iterative marginal optimization） algorithm which directly aims to maximize the penalized rank correlation function is proposed. The effects of the estimating procedure are illustrated by simulation studies.In this paper, we investigate the variable selection problem of the generalized regression models. To estimate the regression parameter, a procedure combining the rank correlation method and the adaptive lasso technique is developed, which is proved to have oracle properties. A modified IMO （iterative marginal optimization） algorithm which directly aims to maximize the penalized rank correlation function is proposed. The effects of the estimating procedure are illustrated by simulation studies.

关 键 词：maximum rank correlation estimation adaptive LASSO oracle properties generalized regression models. 

分 类 号：O212.1[理学—概率论与数理统计] X823[理学—数学]