Abstract
In this work the author investigate a Bayesian inference in lasso censored regression with more flexiblehierarchical prior model. Learning rate parameter proposed to have multinomial distribution which is a priordistribution of the target parameter, so updated hierarchical prior expression developed and Gibbs sample algorithm have independent based on the proposed hierarchical prior model to generate samples from full conditional posterior distribution. Simulation example have conducted to test the performance of the suggested model and real data analysis presented. The result summary of simulation studies and real raw data show that the proposed model perform better than some existing methods .
