Inference in generalized linear model with inequality

Abstract

In this paper we are develop new criteria to estimate parameters of liner regression model under certain linear equality and inequality restrictions on regression coefficients. A new procedure has been suggested for testing inequality hypotheses on regression coefficients of linear model. in this context, the covariance matrix for heteroscedastic disturbances has been specifies and estimated by choosing structure for the error variance. The likelihood ratio test has been discussed for testing linear inequality restrictions on parameters of the generalized linear regression model.