论文信息 - The EM algorithm for mixtures of factor analyzers

The EM algorithm for mixtures of factor analyzers

Factor analysis, a statistical method for modeling the covariance structure of high dimensional data using a small number of latent variables, can be extended by allowing di erent local factor models in di erent regions of the input space. This results in a model which concurrently performs clustering and dimensionality reduction, and can be thought of as a reduced dimension mixture of Gaussians. We present an exact Expectation{Maximization algorithm for tting the parameters of this mixture of factor analyzers.

Geoffrey E. Hinton | Zoubin Ghahramani

[1] Richard O. Duda,et al. Pattern classification and scene analysis , 1974, A Wiley-Interscience publication.

[2] Peter E. Hart,et al. Pattern classification and scene analysis , 1974, A Wiley-Interscience publication.

[3] Dorothy T. Thayer,et al. EM algorithms for ML factor analysis , 1982 .

[4] Brian Everitt,et al. An Introduction to Latent Variable Models , 1984 .

[5] Nanda Kambhatla,et al. Fast Non-Linear Dimension Reduction , 1993, NIPS.

[6] Maurice Milgram,et al. Transformation Invariant Autoassociation with Application to Handwritten Character Recognition , 1994, NIPS.

[7] Geoffrey E. Hinton,et al. Recognizing Handwritten Digits Using Mixtures of Linear Models , 1994, NIPS.

[8] C. Bregler,et al. Surface Learning with Applications to Lipreading Surface Learning with Applications to Lipreading , 1994 .

[9] Geoffrey E. Hinton,et al. Modeling the manifolds of images of handwritten digits , 1997, IEEE Trans. Neural Networks.

[10] Tomaso A. Poggio,et al. Example-Based Learning for View-Based Human Face Detection , 1998, IEEE Trans. Pattern Anal. Mach. Intell..