The Exact Asymptotic Form of Bayesian Generalization Error in Latent Dirichlet Allocation

Abstract

Latent Dirichlet allocation (LDA) obtains essential information from data by using Bayesian inference. It is applied to knowledge discovery via dimension reducing and clustering in many fields. However, its generalization error had not been yet clarified since it is a singular statistical model where there is no one-to-one mapping from parameters to probability distributions. In this paper, we give the exact asymptotic form of its generalization error and marginal likelihood, by theoretical analysis of its learning coefficient using algebraic geometry. The theoretical result shows that the Bayesian generalization error in LDA is expressed in terms of that in matrix factorization and a penalty from the simplex restriction of LDA's parameter region. A numerical experiment is consistent to the theoretical result.

arXiv comment20 pages, 3 figures, 2 tables. Accepted at Neural Networks (Elsevier)

References in corpus (7)

• Asymptotic Equivalence of Bayes Cross Validation and Widely Applicable Information Criterion in Singular Learning Theory
• A Widely Applicable Bayesian Information Criterion
• Stochastic complexity of Bayesian networks
• Upper Bound of Bayesian Generalization Error in Non-negative Matrix Factorization
• Variational Approximation Error in Bayesian Non-negative Matrix Factorization
• Bayesian Generalization Error of Poisson Mixture and Simplex Vandermonde Matrix Type Singularity
• Estimating Real Log Canonical Thresholds