Maintaining Topic Models for Growing Corpora

Felix Kuhr; Magnus Bender; Tanya Braun; Ralf Moller

doi:10.1109/ICSC.2020.00087

Maintaining Topic Models for Growing Corpora

Felix Kuhr, Magnus Bender, Tanya Braun, Ralf Moller

Abstract

A reference library can be described as a corpus of an individual composition of documents. Over time, the corpus might grow because an agent decides to extend its corpus with additional documents, e.g., new publications, or new articles. Existing approaches use topic modelling techniques to compare documents with each other within the same corpus by the documents' topic distribution. However, for new documents, only the text, and no topic distribution is available. Thus, this paper describes three techniques for estimating topic distributions of new unseen documents considering the initial documents in a corpus. Additionally, we present an extensive evaluation about the performance and runtime of the three topic modelling techniques for various scenarios and different sized corpora.

Original language	English
Title of host publication	2020 IEEE 14th International Conference on Semantic Computing (ICSC)
Number of pages	8
Publisher	IEEE
Publication date	02.2020
Pages	451-458
Article number	9031467
ISBN (Print)	978-1-7281-6333-8
ISBN (Electronic)	978-1-7281-6332-1
DOIs	https://doi.org/10.1109/ICSC.2020.00087
Publication status	Published - 02.2020
Event	14th IEEE International Conference on Semantic Computing - San Diego, United States Duration: 03.02.2020 → 05.02.2020 Conference number: 158497

UN SDGs

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

SDG 3 Good Health and Well-being
SDG 4 Quality Education
SDG 9 Industry, Innovation, and Infrastructure
SDG 11 Sustainable Cities and Communities
SDG 12 Responsible Consumption and Production
SDG 14 Life Below Water
SDG 15 Life on Land

Research Areas and Centers

Centers: Center for Artificial Intelligence Luebeck (ZKIL)
Research Area: Intelligent Systems

DFG Research Classification Scheme

409-06 Information Systems, Process and Knowledge Management

Access to Document

10.1109/ICSC.2020.00087

Cite this

@inproceedings{731c642242254fc2af8133b2def7b667,

title = "Maintaining Topic Models for Growing Corpora",

abstract = "A reference library can be described as a corpus of an individual composition of documents. Over time, the corpus might grow because an agent decides to extend its corpus with additional documents, e.g., new publications, or new articles. Existing approaches use topic modelling techniques to compare documents with each other within the same corpus by the documents' topic distribution. However, for new documents, only the text, and no topic distribution is available. Thus, this paper describes three techniques for estimating topic distributions of new unseen documents considering the initial documents in a corpus. Additionally, we present an extensive evaluation about the performance and runtime of the three topic modelling techniques for various scenarios and different sized corpora.",

author = "Felix Kuhr and Magnus Bender and Tanya Braun and Ralf Moller",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE. Copyright: Copyright 2020 Elsevier B.V., All rights reserved.; 14th IEEE International Conference on Semantic Computing , ICSC 2020 ; Conference date: 03-02-2020 Through 05-02-2020",

year = "2020",

month = feb,

doi = "10.1109/ICSC.2020.00087",

language = "English",

isbn = "978-1-7281-6333-8",

series = "IEEE International Conference on Semantic Computing",

publisher = "IEEE",

pages = "451--458",

booktitle = "2020 IEEE 14th International Conference on Semantic Computing (ICSC)",

}

Kuhr, F, Bender, M, Braun, T & Moller, R 2020, Maintaining Topic Models for Growing Corpora. in 2020 IEEE 14th International Conference on Semantic Computing (ICSC)., 9031467, IEEE International Conference on Semantic Computing, IEEE, pp. 451-458, 14th IEEE International Conference on Semantic Computing , San Diego, California, United States, 03.02.20. https://doi.org/10.1109/ICSC.2020.00087

TY - GEN

T1 - Maintaining Topic Models for Growing Corpora

AU - Kuhr, Felix

AU - Bender, Magnus

AU - Braun, Tanya

AU - Moller, Ralf

N1 - Conference code: 158497

PY - 2020/2

Y1 - 2020/2

N2 - A reference library can be described as a corpus of an individual composition of documents. Over time, the corpus might grow because an agent decides to extend its corpus with additional documents, e.g., new publications, or new articles. Existing approaches use topic modelling techniques to compare documents with each other within the same corpus by the documents' topic distribution. However, for new documents, only the text, and no topic distribution is available. Thus, this paper describes three techniques for estimating topic distributions of new unseen documents considering the initial documents in a corpus. Additionally, we present an extensive evaluation about the performance and runtime of the three topic modelling techniques for various scenarios and different sized corpora.

AB - A reference library can be described as a corpus of an individual composition of documents. Over time, the corpus might grow because an agent decides to extend its corpus with additional documents, e.g., new publications, or new articles. Existing approaches use topic modelling techniques to compare documents with each other within the same corpus by the documents' topic distribution. However, for new documents, only the text, and no topic distribution is available. Thus, this paper describes three techniques for estimating topic distributions of new unseen documents considering the initial documents in a corpus. Additionally, we present an extensive evaluation about the performance and runtime of the three topic modelling techniques for various scenarios and different sized corpora.

UR - https://www.scopus.com/pages/publications/85083444821

UR - https://www.mendeley.com/catalogue/17b3e211-af38-31ac-aa89-8468c3fddbe9/

U2 - 10.1109/ICSC.2020.00087

DO - 10.1109/ICSC.2020.00087

M3 - Conference contribution

SN - 978-1-7281-6333-8

T3 - IEEE International Conference on Semantic Computing

SP - 451

EP - 458

BT - 2020 IEEE 14th International Conference on Semantic Computing (ICSC)

PB - IEEE

T2 - 14th IEEE International Conference on Semantic Computing

Y2 - 3 February 2020 through 5 February 2020

ER -

Maintaining Topic Models for Growing Corpora

Abstract

UN SDGs

Research Areas and Centers

DFG Research Classification Scheme

Access to Document

Other files and links

Fingerprint

Cite this