United we stand: Using multiple strategies for topic labeling

Abstract : Topic labeling aims at providing a sound, possibly multi-words, label that depicts a topic drawn from a topic model. This is of the utmost practical interest in order to quickly grasp a topic informa-tional content-the usual ranked list of words that maximizes a topic presents limitations for this task. In this paper, we introduce three new unsupervised n-gram topic labelers that achieve comparable results than the existing unsupervised topic labelers but following different assumptions. We demonstrate that combining topic labelers-even only two-makes it possible to target a 64% improvement with respect to single topic labeler approaches and therefore opens research in that direction. Finally, we introduce a fourth topic labeler that extracts representative sentences, using Dirichlet smoothing to add contextual information. This sentence-based labeler provides strong surrogate candidates when n-gram topic labelers fall short on providing relevant labels, leading up to 94% topic covering.
Document type :
Conference papers
Complete list of metadatas

Cited literature [17 references]  Display  Hide  Download

https://hal-lirmm.ccsd.cnrs.fr/lirmm-01910614
Contributor : Pascal Poncelet <>
Submitted on : Thursday, November 1, 2018 - 3:23:58 PM
Last modification on : Wednesday, November 20, 2019 - 2:49:23 AM
Long-term archiving on: Saturday, February 2, 2019 - 2:00:06 PM

File

NLDB_Julien.pdf
Files produced by the author(s)

Identifiers

Citation

Antoine Gourru, Julien Velcin, Mathieu Roche, Christophe Gravier, Pascal Poncelet. United we stand: Using multiple strategies for topic labeling. NLDB: Natural Language Processing and Information Systems, Jun 2018, Paris, France. pp.352-363, ⟨10.1007/978-3-319-91947-8_37⟩. ⟨lirmm-01910614⟩

Share

Metrics

Record views

175

Files downloads

230