Open AccessAn application of two sided power distribution in Bayesian analysis of paired comparison of relative importance of predictors in linear regression models
Author(s) -
Xiaoyin Wang
Publication year - 2015
Publication title -
international journal of advanced statistics and probability
Language(s) - English
Resource type - Journals
ISSN - 2307-9045
DOI - 10.14419/ijasp.v3i2.4573
Subject(s) - bayesian probability , logistic regression , statistics , econometrics , function (biology) , linear regression , logit , regression analysis , computer science , statistical model , regression , predictive power , mathematics , evolutionary biology , biology , philosophy , epistemology
The purpose of determining the relative importance of predictors is to expose the extent of the individual contribution of a predictor in the presence of other predictors within a selected model. The goal of this article is to expand the current research practice by developing a statistical paired comparison model with Two Sided Power (TSP) link function in the Bayesian framework to evaluate the relative importance of each predictor in a multiple regression model. Results from simulation studies and empirical example reveal that the proposed Two Sided Power link function provides similar conclusions as the commonly used logit link function, but has more advantages from both practical and computational perspectives.