Classification Performance Comparison between SVM and RVM

초록

The Relevance Vector Machine or RVM is a Bayesian sparse kernel technique for regression and classification that shares many of the characteristics of the SVM whilst avoiding its principal limitations. In this paper we compare the performance of RVM with the SVM in classification. SVM is known to be one of the top competitors among several machine learning tools. Using several simulated data set and real data set we compare the classification performance between SVM and RVM to find out relative merits of each method. The principal disadvantage of the RVM is the relatively long training time compared with the SVM. This is offset, however, by the avoidance of cross-validation runs to set the model complexity parameters. Furthermore, because it yields sparser models, the computation time on test points is typically much less.