IEcons: A New Consensus Approach Using Multi-Text Representations for Clustering Task - Laboratoire d’Informatique, Systèmes, Traitement de l’Information et de la Connaissance
Communication Dans Un Congrès Année : 2024

IEcons: A New Consensus Approach Using Multi-Text Representations for Clustering Task

Karima Boutalbi
Rafika Boutalbi
Hervé Verjus
Kave Salamatian
Olivier Le Van
  • Fonction : Auteur
  • PersonId : 1429129

Résumé

Today we are able to generate a large set of text representations from the simple Bag-of-word (BOW) to the recent transformers capturing the semantic and the contextual text meaning. It was proven that there is no best text representation for text clustering task. Consequently, some works combined text representations using a consensus clustering approach. Two consensus approach types exist, namely explicit and implicit consensus. In the explicit consensus, also known as ensemble clustering, the consensus function is applied a posterior after obtaining cluster labels from each text representation clustering allowing to capture global mutual information between the partitions of all text representations. On the other hand, implicit consensus uses tensor clustering to optimize the clustering consensus partition that deals with similarity matrices of text representations. In this paper, we propose a new consensus text clustering algorithm named IEcons (Implicit-Explicit consensus) that optimizes explicit and implicit consensus clustering simultaneously through text embeddings and tensor representation of texts through similarity matrices. We compare our algorithm with others from the literature on five different textual datasets using several algorithm performance criteria. The comparison results reveal that our algorithm best suits most situations.

Fichier principal
Vignette du fichier
CIKM_IEcons__An_Approach_Using_Multi_Text_Representations_for_Clustering_Task (2).pdf (646.07 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04741799 , version 1 (17-10-2024)

Identifiants

  • HAL Id : hal-04741799 , version 1

Citer

Karima Boutalbi, Rafika Boutalbi, Hervé Verjus, Kave Salamatian, David Telisson, et al.. IEcons: A New Consensus Approach Using Multi-Text Representations for Clustering Task. CIKM24: 33rd ACM International Conference on Information and Knowledge Management, Oct 2024, BOISE, United States. pp.613 - 616. ⟨hal-04741799⟩

Collections

UNIV-SAVOIE LISTIC
0 Consultations
0 Téléchargements

Partager

More