论文信息 - Effective Dimensions of Partially Observed Polytrees

Effective Dimensions of Partially Observed Polytrees

Model complexity is an important factor to consider when selecting among graphical models. When all variables are observed, the complexity of a model can be measured by its standard dimension, i.e. the number of independent parameters. When latent variables are present, however, the standard dimension might no longer be appropriate. Instead, an effective dimension should be used [5]. Zhang & Kocka [13] showed how to compute the effective dimensions of partially observed trees. In this paper we solve the same problem for partially observed polytrees.

Tomas Kocka | Nevin Lianwen Zhang

[1] Tomas Kocka,et al. Effective Dimensions of Hierarchical Latent Class Models , 2011, J. Artif. Intell. Res..

[2] David Heckerman,et al. Asymptotic Model Selection for Directed Networks with Hidden Variables , 1996, UAI.

[3] Peter C. Cheeseman,et al. Bayesian Classification (AutoClass): Theory and Results , 1996, Advances in Knowledge Discovery and Data Mining.

[4] Jim Q. Smith,et al. On the Geometry of Bayesian Graphical Models with Hidden Variables , 1998, UAI.

[5] Tomas Kocka,et al. Dimension Correction for Hierarchical Latent Class Models , 2002, UAI.

[6] H. Akaike. A new look at the statistical model identification , 1974 .

[7] Byron J. T. Morgan,et al. Deficiency of parameter‐redundant models , 2001 .

[8] Gregory F. Cooper,et al. A Bayesian method for the induction of probabilistic networks from data , 1992, Machine Learning.

[9] G. Schwarz. Estimating the Dimension of a Model , 1978 .

[10] Jim Q. Smith,et al. Geometry, moments and Bayesian networks with hidden variables , 1999, AISTATS.