Sparse Bayesian Unsupervised Learning
Machine Learning
2014-02-03 v1
Abstract
This paper is about variable selection, clustering and estimation in an unsupervised high-dimensional setting. Our approach is based on fitting constrained Gaussian mixture models, where we learn the number of clusters and the set of relevant variables using a generalized Bayesian posterior with a sparsity inducing prior. We prove a sparsity oracle inequality which shows that this procedure selects the optimal parameters and . This procedure is implemented using a Metropolis-Hastings algorithm, based on a clustering-oriented greedy proposal, which makes the convergence to the posterior very fast.
Cite
@article{arxiv.1401.8017,
title = {Sparse Bayesian Unsupervised Learning},
author = {Stephane Gaiffas and Bertrand Michel},
journal= {arXiv preprint arXiv:1401.8017},
year = {2014}
}