Determination of the Number of Topics Intrinsically: Is It Possible?

arXiv:2406.10402 · cs.CL, cs.IR, math.PR · Submitted 2024-06-14 · Read on arXiv

Victor Bulatov, Vasiliy Alekseev, Konstantin Vorontsov

cs.CL, cs.IR, math.PR

Submitted: 2024-06-14

Updated: 2026-08-21

Comments: This is the first full draft version of the article. The camera-ready version was accepted at the 11th International Conference on Analysis of Images, Social Networks and Texts (AIST 2023). Presented on September 30, 2023. Expected to be published in the conference proceedings, as part of the Communications in Computer and Information Science series (CCIS, Vol. 1905)

Journal ref: Recent Trends in Analysis of Images, Social Networks and Texts. AIST 2023. Communications in Computer and Information Science (CCIS), vol 1905. Springer, Cham

DOI: 10.1007/978-3-031-67008-4_1

Code: https://github.com/nikita-moor/ldatuninghttps:

License: http://creativecommons.org/licenses/by/4.0/

The gist: The number of topics might be the most important parameter of a topic model.

Terminology

Abstract

The number of topics might be the most important parameter of a topic model. The topic modelling community has developed a set of various procedures to estimate the number of topics in a dataset, but there has not yet been a sufficiently complete comparison of existing practices. This study attempts to partially fill this gap by investigating the performance of various methods applied to several topic models on a number of publicly available corpora. Further analysis demonstrates that intrinsic methods are far from being reliable and accurate tools. The number of topics is shown to be a method- and a model-dependent quantity, as opposed to being an absolute property of a particular corpus. We conclude that other methods for dealing with this problem should be developed and suggest some promising directions for further research.

Sources

Related papers