Magnus Rattray et al 2009 J. Phys.: Conf. Ser. 197 012002 doi:10.1088/1742-6596/197/1/012002
Magnus Rattray1, Oliver Stegle2,3, Kevin Sharp1 and John Winn4
Show affiliationsBayesian sparse factor analysis has many applications; for example, it has been applied to the problem of inferring a sparse regulatory network from gene expression data. We describe a number of inference algorithms for Bayesian sparse factor analysis using a slab and spike mixture prior. These include well-established Markov chain Monte Carlo (MCMC) and variational Bayes (VB) algorithms as well as a novel hybrid of VB and Expectation Propagation (EP). For the case of a single latent factor we derive a theory for learning performance using the replica method. We compare the MCMC and VB/EP algorithm results with simulated data to the theoretical prediction. The results for MCMC agree closely with the theory as expected. Results for VB/EP are slightly sub-optimal but show that the new algorithm is effective for sparse inference. In large-scale problems MCMC is infeasible due to computational limitations and the VB/EP algorithm then provides a very useful computationally efficient alternative.
87.80.-y Biophysical techniques (research methods)
87.15.B- Structure of biomolecules
Issue 1 (2009)
Magnus Rattray et al 2009 J. Phys.: Conf. Ser. 197 012002
E G Thorsteinsson and J T Gudmundsson 2010 Plasma Sources Sci. Technol. 19 015001
T Abe et al 2009 J. Phys.: Conf. Ser. 191 012023
E A Brener et al 2009 J. Phys.: Condens. Matter 21 464106
S Desai et al 2008 Class. Quantum Grav. 25 184024
A Ruiz et al 2007 Nanotechnology 18 505306
Dongmin Gang et al JHEP11(2009)024
Akio Kotani et al 2009 J. Phys.: Conf. Ser. 190 012013
Qiu-Bo Fan and Ling Zhou 2009 J. Phys. B: At. Mol. Opt. Phys. 42 145509
Chien-Sheng Kuo et al 2007 Nanotechnology 18 465607