تقدير بيز التقريبي لمعلمات نموذج انحدار t المبتور الخطي البسيط

Abstract EN

Researchers treat many phenomena as having a non-truncated probability distribution, but in fact belong to truncated distributions.

The estimators of the parameters by using the truncated distributions more efficiently than non-truncated distributions.

This phenomenon appears in linear regression models.

For this reason, this study is interested in estimating the parameters of simple linear regression truncated t model when the response variable follows t truncated distribution from two sides by using bayes method.

Because a cumulative function can not be treated as a fixed quantity, it is difficult to find the posterior marginal probability distributions for each parameter of the model.

Therefore, a lundley's approximation that deals with the joint posterior probability density function of the model parameters has been used to obtain expected Marginal posterior.