Regression Based on Sparse Bayesian Learning and the Applications in Electric Systems

This paper introduces a general Bayesian framework for obtaining sparse solutions to regression predicting, and the practical model 'relevance vector machine' (RVM) by Michael E. Tipping. As a bran-new thought of probabilistic learning model, it offers the superior level of generalization accuracy and a number of additional advantages comparable with the popular and state-of-the-art 'support vector machine' (SVM). Utilize the advantages of the RVM, it can be applied in sorts of practical engineering fields and gain the special benefits. In this paper we also give the perspective of the model in electric systems regression implementations. A short-term electricity load prediction model is presented as an example.