Analyzing relevance vector machines using a single penalty approach

Relevance vector machine (RVM) is a popular sparse Bayesian learning model typically used for prediction. Recently it has been shown that improper priors assumed on multiple penalty parameters in RVM may lead to an posterior. Currently the literature, sufficient conditions posterior propriety of do not allow over parameters. In this article, we propose single relevance (SPRVM) which are replaced by and consider semi-Bayesian approach fitting SPRVM. The necessary SPRVM more liberal than those several parameter. Additionally, also prove geometric ergodicity Gibbs sampler analyze hence can estimate asymptotic standard errors associated with Monte Carlo means predictive distribution. Such error cannot be computed case RVM, since rate convergence known. performance compared analyzing two simulation examples three real life datasets.