Bayesian parameter estimation in addiction model

Document Type : Research Paper


Department of Mathematics, College of Science, University of Baghdad, Iraq


In this paper, we investigated the performance of Bayesian Computational methods for estimating the parameters of the multinomial Logistic regression model. We discussed two of the most common Bayesian computational algorithms: the Random walk Metropolis-Hastings (RWM) and Slice algorithms and their application to estimating the parameters of the addiction model as well as comparing the performance of these algorithms using the mean square error  (MSE) criterion. The results revealed that the performance of the algorithms is excellent, with a slight superiority to the RWM algorithm.