Relevance Vector Machines and Eponine

Konstantin Tretyakov
Institute of Computer Science, University of Tartu

Bioinformatics of Gene Regulation seminar, MTAT.03.192, Fall 2005

Abstract

Relevance vector machines (RVM) is a family of machine learning methods, introduced by Tipping, that represent a bayesian approach to the training of general linear models (GLM). RVM-s are reported to be able to effectively produce convenient sparse representations of data thus competing with the popular support vector machines. When used with a suitable set of basis functions, RVM-s can be applied to the analysis of genome sequence data. The Eponine models provide such sets of functions. This article gives a brief (and probably slightly simplistic) overview of the application of the RVM with Eponine to genome sequence analysis, a work originally described by Down.

Download

• Article (pdf)
• Presentation slides (pdf)

Page last updated: 23.12.2005