Bayes Hypothesis Testing of Multivariate Linear Regression Model

摘要

In this paper,the unknown parameter β'S posterior distribution in the linear model Y=Xβ+e is discussed,where e～Nn(0,σ2In),then an ratio of β'S maximum probability distribution functions in Rp and one of its subsets is defined,finally,two special situations are tested:①Hypothesis testing of regression equation,H0 ∶β1 =β2 βp-1 ②Hypothesis testing of regression coefficient,H1 ∶ βi =0(i =1,2,…,p-1).