Hierarchical Dirichlet scaling process

DOI10.1007/S10994-016-5621-5MaRDI QIDQ2014581zbMATH OpenOpenAlexDBLPWikidataFDO

Authors Dongwoo Kim, Alice Oh

Publication date 25 August 2017

Published in Machine Learning (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/1404.1282

zbMATH Keywords

topic modelingDirichlet processhierarchical Dirichlet processBayesian nonparametric mixed membership model

Mathematics Subject Classification ID

Nonparametric estimation (62G05) Classification and discrimination; cluster analysis (statistical aspects) (62H30) Applications of statistics to economics (62P20)

Abstract: We present the extit{hierarchical Dirichlet scaling process} (HDSP), a Bayesian nonparametric mixed membership model. The HDSP generalizes the hierarchical Dirichlet process (HDP) to model the correlation structure between metadata in the corpus and mixture components. We construct the HDSP based on the normalized gamma representation of the Dirichlet process, and this construction allows incorporating a scaling function that controls the membership probabilities of the mixture components. We develop two scaling methods to demonstrate that different modeling assumptions can be expressed in the HDSP. We also derive the corresponding approximate posterior inference algorithms using variational Bayes. Through experiments on datasets of newswire, medical journal articles, conference proceedings, and product reviews, we show that the HDSP results in a better predictive performance than labeled LDA, partially labeled LDA, and author topic model and a better negative review classification performance than the supervised topic model and SVM.

Recommendations

• Hierarchical Dirichlet Processes
• Dynamic hierarchical Dirichlet processes topic model using the power prior approach
• Hierarchical topic modeling with nested hierarchical Dirichlet process
• 10.1162/jmlr.2003.3.4-5.993
• The discrete infinite logistic normal distribution

Cites work

• scientific article; zbMATH DE number 6377992 (Why is no real title available?)
• scientific article; zbMATH DE number 774881 (Why is no real title available?)
• 10.1162/jmlr.2003.3.4-5.993
• A correlated topic model of science
• An introduction to variational methods for graphical models
• Bayesian Density Estimation and Inference Using Mixtures
• Bayesian Nonparametric Spatial Modeling With Dirichlet Process Mixing
• Generalized Spatial Dirichlet Process Models
• Graphical models, exponential families, and variational inference
• Hierarchical Dirichlet Processes
• Kernel stick-breaking processes
• Logistic stick-breaking process
• Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis
• Order-Based Dependent Dirichlet Processes
• Pattern recognition and machine learning.
• The discrete infinite logistic normal distribution
• Variational inference for Dirichlet process mixtures

Cited in

(6)

• Bayesian hierarchical dictionary learning
• Cascade model with Dirichlet process for analyzing multiple dyadic matrices
• The discrete infinite logistic normal distribution
• Hierarchical topic modeling with nested hierarchical Dirichlet process
• Shifted-Scaled Dirichlet-Based Hierarchical Dirichlet Process Hidden Markov Models with Variational Inference Learning
• Dynamic hierarchical Dirichlet processes topic model using the power prior approach

Describes a project that uses

Uses Software

• PRMLT
• Medlda

This page was built for publication: Hierarchical Dirichlet scaling process

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2014581)