Variable selection in multivariate multiple regression

Variyath, Asokan Mulayath and Brobbey, Anita (2020) Variable selection in multivariate multiple regression. PLoS ONE, 15 (7). ISSN 1932-6203

[img] [English] PDF - Published Version
Available under License Creative Commons Attribution Non-commercial.

Download (1MB)


Introduction In many practical situations, we are interested in the effect of covariates on correlated multiple responses. In this paper, we focus on estimation and variable selection in multi-response multiple regression models. Correlation among the response variables must be modeled for valid inference. Method We used an extension of the generalized estimating equation (GEE) methodology to simultaneously analyze binary, count, and continuous outcomes with nonlinear functions. Variable selection plays an important role in modeling correlated responses because of the large number of model parameters that must be estimated. We propose a penalized-likelihood approach based on the extended GEEs for simultaneous parameter estimation and variable selection. Results and conclusions We conducted a series of Monte Carlo simulations to investigate the performance of our method, considering different sample sizes and numbers of response variables. The results showed that our method works well compared to treating the responses as uncorrelated. We recommend using an unstructured correlation model with the Bayesian information criterion (BIC) to select the tuning parameters. We demonstrated our method using data from a concrete slump test.

Item Type: Article
Item ID: 14887
Additional Information: Memorial University Open Access Author's Fund
Department(s): Science, Faculty of > Mathematics and Statistics
Date: 17 July 2020
Date Type: Publication
Digital Object Identifier (DOI): journal.pone.0236067
Related URLs:

Actions (login required)

View Item View Item


Downloads per month over the past year

View more statistics