Parallel Dynamic Topic Modeling via Evolving Topic Adjustment and Term Weighting Scheme

Hongyu JIANG; Zhiqi LEI; Yanghui RAO; Haoran XIE; Fu Lee WANG

doi:10.1016/j.ins.2021.11.060

Parallel Dynamic Topic Modeling via Evolving Topic Adjustment and Term Weighting Scheme

Hongyu JIANG, Zhiqi LEI, Yanghui RAO^*, Haoran XIE, Fu Lee WANG

^*Corresponding author for this work

Department of Computing and Decision Sciences

Research output: Journal Publications › Journal Article (refereed) › peer-review

1 Citation (Scopus)

Abstract

The parallel Hierarchical Dirichlet Process (pHDP) is an efficient topic model which explores the equivalence of the generation process between Hierarchical Dirichlet Process (HDP) and Gamma-Gamma-Poisson Process (G2PP), in order to achieve parallelism at the topic level. Unfortunately, pHDP loses the non-parametric feature of HDP, i.e., the number of topics in pHDP is predetermined and fixed. Furthermore, under the bootstrap structure of pHDP, the topic-indiscriminate words are of high probabilities to be assigned to different topics, resulting in poor qualities of the extracted topics. To achieve parallelism without sacrificing the non-parametric feature of HDP, in addition to improve the quality of extracted topics, we propose a parallel dynamic topic model by developing an adjustment mechanism of evolving topics and reducing the sampling probabilities of topic-indiscriminate words. Both supervised and unsupervised experiments on benchmark datasets show the competitive performance of our model.

Original language	English
Pages (from-to)	176-193
Number of pages	18
Journal	Information Sciences
Volume	585
Early online date	26 Nov 2021
DOIs	https://doi.org/10.1016/j.ins.2021.11.060
Publication status	Published - Mar 2022

Bibliographical note

Funding Information:
The authors are thankful to the reviewers for their constructive comments and suggestions. The research described in this paper was supported by the National Natural Science Foundation of China (61972426), Guangdong Basic and Applied Basic Research Foundation (2020A1515010536), the Direct Grant (DR21A5) and the Faculty Research Grants (DB21B6 and DB21A9) of Lingnan University, Hong Kong, and a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (UGC/FDS16/E01/19).

Publisher Copyright:
© 2021 Elsevier Inc.

Keywords

Dynamic topic model
Parallel gibbs sampling
Term weighting scheme

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1016/j.ins.2021.11.060

Cite this

@article{e4b46ff9f57147f2b0826eb9d9d34897,

title = "Parallel Dynamic Topic Modeling via Evolving Topic Adjustment and Term Weighting Scheme",

abstract = "The parallel Hierarchical Dirichlet Process (pHDP) is an efficient topic model which explores the equivalence of the generation process between Hierarchical Dirichlet Process (HDP) and Gamma-Gamma-Poisson Process (G2PP), in order to achieve parallelism at the topic level. Unfortunately, pHDP loses the non-parametric feature of HDP, i.e., the number of topics in pHDP is predetermined and fixed. Furthermore, under the bootstrap structure of pHDP, the topic-indiscriminate words are of high probabilities to be assigned to different topics, resulting in poor qualities of the extracted topics. To achieve parallelism without sacrificing the non-parametric feature of HDP, in addition to improve the quality of extracted topics, we propose a parallel dynamic topic model by developing an adjustment mechanism of evolving topics and reducing the sampling probabilities of topic-indiscriminate words. Both supervised and unsupervised experiments on benchmark datasets show the competitive performance of our model.",

keywords = "Dynamic topic model, Parallel gibbs sampling, Term weighting scheme",

author = "Hongyu JIANG and Zhiqi LEI and Yanghui RAO and Haoran XIE and WANG, {Fu Lee}",

note = "Funding Information: The authors are thankful to the reviewers for their constructive comments and suggestions. The research described in this paper was supported by the National Natural Science Foundation of China (61972426), Guangdong Basic and Applied Basic Research Foundation (2020A1515010536), the Direct Grant (DR21A5) and the Faculty Research Grants (DB21B6 and DB21A9) of Lingnan University, Hong Kong, and a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (UGC/FDS16/E01/19). Publisher Copyright: {\textcopyright} 2021 Elsevier Inc.",

year = "2022",

month = mar,

doi = "10.1016/j.ins.2021.11.060",

language = "English",

volume = "585",

pages = "176--193",

journal = "Information Sciences",

issn = "0020-0255",

publisher = "Elsevier Inc.",

}

TY - JOUR

T1 - Parallel Dynamic Topic Modeling via Evolving Topic Adjustment and Term Weighting Scheme

AU - JIANG, Hongyu

AU - LEI, Zhiqi

AU - RAO, Yanghui

AU - XIE, Haoran

AU - WANG, Fu Lee

N1 - Funding Information: The authors are thankful to the reviewers for their constructive comments and suggestions. The research described in this paper was supported by the National Natural Science Foundation of China (61972426), Guangdong Basic and Applied Basic Research Foundation (2020A1515010536), the Direct Grant (DR21A5) and the Faculty Research Grants (DB21B6 and DB21A9) of Lingnan University, Hong Kong, and a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (UGC/FDS16/E01/19). Publisher Copyright: © 2021 Elsevier Inc.

PY - 2022/3

Y1 - 2022/3

N2 - The parallel Hierarchical Dirichlet Process (pHDP) is an efficient topic model which explores the equivalence of the generation process between Hierarchical Dirichlet Process (HDP) and Gamma-Gamma-Poisson Process (G2PP), in order to achieve parallelism at the topic level. Unfortunately, pHDP loses the non-parametric feature of HDP, i.e., the number of topics in pHDP is predetermined and fixed. Furthermore, under the bootstrap structure of pHDP, the topic-indiscriminate words are of high probabilities to be assigned to different topics, resulting in poor qualities of the extracted topics. To achieve parallelism without sacrificing the non-parametric feature of HDP, in addition to improve the quality of extracted topics, we propose a parallel dynamic topic model by developing an adjustment mechanism of evolving topics and reducing the sampling probabilities of topic-indiscriminate words. Both supervised and unsupervised experiments on benchmark datasets show the competitive performance of our model.

AB - The parallel Hierarchical Dirichlet Process (pHDP) is an efficient topic model which explores the equivalence of the generation process between Hierarchical Dirichlet Process (HDP) and Gamma-Gamma-Poisson Process (G2PP), in order to achieve parallelism at the topic level. Unfortunately, pHDP loses the non-parametric feature of HDP, i.e., the number of topics in pHDP is predetermined and fixed. Furthermore, under the bootstrap structure of pHDP, the topic-indiscriminate words are of high probabilities to be assigned to different topics, resulting in poor qualities of the extracted topics. To achieve parallelism without sacrificing the non-parametric feature of HDP, in addition to improve the quality of extracted topics, we propose a parallel dynamic topic model by developing an adjustment mechanism of evolving topics and reducing the sampling probabilities of topic-indiscriminate words. Both supervised and unsupervised experiments on benchmark datasets show the competitive performance of our model.

KW - Dynamic topic model

KW - Parallel gibbs sampling

KW - Term weighting scheme

UR - http://www.scopus.com/inward/record.url?scp=85120378789&partnerID=8YFLogxK

U2 - 10.1016/j.ins.2021.11.060

DO - 10.1016/j.ins.2021.11.060

M3 - Journal Article (refereed)

SN - 0020-0255

VL - 585

SP - 176

EP - 193

JO - Information Sciences

JF - Information Sciences

ER -

Parallel Dynamic Topic Modeling via Evolving Topic Adjustment and Term Weighting Scheme

Abstract

Bibliographical note

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

A Label Extension Schema for Improved Text Emotion Classification

Facilitate Tree-Structured Topic Modeling via Nonparametric Neural Inference

Cite this

Parallel Dynamic Topic Modeling via Evolving Topic Adjustment and Term Weighting Scheme

Abstract

Bibliographical note

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Projects

A Label Extension Schema for Improved Text Emotion Classification

Facilitate Tree-Structured Topic Modeling via Nonparametric Neural Inference

Cite this