Skip to main content

Fast Moment Estimation for Generalized Latent Dirichlet Models

Author(s): Zhao, Shiwen; Engelhardt, Barbara E; Mukherjee, Sayan; Dunson, David B

To refer to this page use:
Abstract: We develop a generalized method of moments (GMM) approach for fast parameter estimation in a new class of Dirichlet latent variable models with mixed data types. Parameter estimation via GMM has computational and statistical advantages over alternative methods, such as expectation maximization, variational inference, and Markov chain Monte Carlo. A key computational advantage of our method, Moment Estimation for latent Dirichlet models (MELD), is that parameter estimation does not require instantiation of the latent variables. Moreover, performance is agnostic to distributional assumptions of the observations. We derive population moment conditions after marginalizing out the sample-specific Dirichlet latent variables. The moment conditions only depend on component mean parameters. We illustrate the utility of our approach on simulated data, comparing results from MELD to alternative methods, and we show the promise of our approach through the application to several datasets. Supplementary materials for this article are available online.
Publication Date: 2018
Citation: Zhao, Shiwen, Barbara E. Engelhardt, Sayan Mukherjee, and David B. Dunson. "Fast Moment Estimation for Generalized Latent Dirichlet Models." Journal of the American Statistical Association 113, no. 524 (2018): pp. 1528-1540. doi:10.1080/01621459.2017.1341839
DOI: 10.1080/01621459.2017.1341839
ISSN: 0162-1459
EISSN: 1537-274X
Pages: 1528 - 1540
Type of Material: Journal Article
Journal/Proceeding Title: Journal of the American Statistical Association
Version: Author's manuscript

Items in OAR@Princeton are protected by copyright, with all rights reserved, unless otherwise indicated.