Robust Probabilistic Projections
(2006) 23rd International Conference on Machine Learning (ICML 2006) — Location: Pittsburgh (PA/USA) (26.June.2006)
Files
127-Robustprobabilisticprojections.pdf
Restricted Access - Adobe PDF
- 286.89 KB
Details
- Authors
- Abstract
- Principal components and canonical correlations are at the root of many exploratory data mining techniques and provide standard pre-processing tools in machine learning. Lately, probabilistic reformulations of these methods have been proposed (Roweis, 1998; Tipping & Bishop, 1999b; Bach & Jordan, 2005). They are based on a Gaussian density model and are therefore, like their non-probabilistic counterpart, very sensitive to atypical observations. In this paper, we introduce robust probabilistic principal component analysis and robust probabilistic canonical correlation analysis. Both are based on a Student-t density model. The resulting probabilistic reformulations are more suitable in practice as they handle outliers in a natural way. We compute maximum ikelihood estimates of the parameters by means of the EM algorithm.
- Affiliations
Citations
Archambeau, C., Delannay, N., & Verleysen, M. (2006). Robust Probabilistic Projections. Proceedings of the 23rd International Conference on Machine Learning (ICML 2006), p. 33-40. https://hdl.handle.net/2078.5/254111