Generative linear mixture modelling

Lawson, Antony; Einbeck, Jochen

Abstract

For multivariate data with a low–dimensional latent structure, a novel approach to linear dimension reduction based on Gaussian mixture models is pro- posed. A generative model is assumed for the data, where the mixture centres (or ‘mass points’) are positioned along lines or planes spanned through the data cloud. All involved parameters are estimated simultaneously through the EM al- gorithm, requiring an additional iteration within each M-step. Data points can be projected onto the low–dimensional space by taking the posterior mean over the estimated mass points. The compressed data can then be used for further pro- cessing, for instance as a low–dimensional predictor in a multivariate regression problem.

Citation

Lawson, A., & Einbeck, J. (2012, December). Generative linear mixture modelling. Presented at International workshop on statistical modelling., Prague

• Presentation Conference Type: Conference Paper (published)
• Conference Name: International workshop on statistical modelling.
• Publication Date: Jan 1, 2012
• Deposit Date: Sep 24, 2012
• Publicly Available Date: Jun 19, 2013
• Volume: 2
• Pages: 595-600
• Series Title: Proceedings of the international workshop on statistical modelling.
• Book Title: 27th International Workshop on Statistical Modelling, 16-20 July 2012, Prague, Czech Republic ; proceedings.
• Keywords: EM, Dimension reduction, Mixture modelling.
• Public URL: https://durham-repository.worktribe.com/output/1157680
• Publisher URL: http://www.statmod.org/workshops_archive_proceedings_2012.html
• Additional Information: http://www.maths.dur.ac.uk/~dma0je/Papers/lawson_einbeck_iwsm2012.pdf