Abstract

The paper provides a multivariate extension of generalized linear model for polytomous data and considers the logistic regression model as a special case. Complete Bayes analysis of polytomous data assuming logistic regression model is provided using appropriate non-informative priors for the parameters. Since the resulting posteriors analysis is quite complex, appropriate Markov chain Monte Carlo algorithm has been developed for the same. Results are illustrated on the basis of a real data example related to biliary acid constituents of the patients having gallbladder diseases.

Author: Richa Srivastava , S. K. Upadhyay and V. K. Shukla

Received on: March, 2015

Accepted on: June, 2016