Robust variable selection in finite mixture of regression models using the t distribution. Issue 21 (2nd November 2019)

• Record Type: Journal Article
• Authors: Dai, Lin; Yin, Junhui; Xie, Zhengfen; Wu, Liucang
• Is Part Of: Communications in statistics. Volume 48:Issue 21(2019)
• Page Start: 5370
• Page End: 5386
• Publication Date: 2019-11-02
• Subjects: EM algorithm -- Hard -- LASSO -- mixture model -- SCAD -- t distribution -- variable selection; 62F35 -- 62H30 -- 62J07; Mathematical statistics -- Periodicals; Mathematics; Statistics; 519.2
• DOI: 10.1080/03610926.2018.1513143
• Languages: English
• ISSNs: 0361-0926

Abstract

Abstract: Variable selection in finite mixture of regression (FMR) models is frequently used in statistical modeling. The majority of applications of variable selection in FMR models use a normal distribution for regression error. Such assumptions are unsuitable for a set of data containing a group or groups of observations with heavy tails and outliers. In this paper, we introduce a robust variable selection procedure for FMR models using the t distribution. With appropriate selection of the tuning parameters, the consistency and the oracle property of the regularized estimators are established. To estimate the parameters of the model, we develop an EM algorithm for numerical computations and a method for selecting tuning parameters adaptively. The parameter estimation performance of the proposed model is evaluated through simulation studies. The application of the proposed model is illustrated by analyzing a real data set.