Evaluarea şi sumarizarea automată a conversaţiilor chat

Abstract : With the continuous evolution of collaborative environments, the need for automatic analysis and participant assessment in Instant Messenger discussions (chats) has become more and more important. Moreover, a key element in a grading and evaluation system for participants and utterances is summarization. In order to achieve the previous goals, a series of factors specific to natural language processing (including semantic analysis by means of Latent Semantic Analysis - LSA) and data-mining were taken into consideration. In order to achieve a thorough and multi-perspective approach for assessing participants, the system includes facilities derived from Page’s essay grading techniques, readability assessment and metrics specific to social networks analysis. A golden standard was used for assessing the relevance of the obtained results and for validation purposes of the proposed method
Type de document :
Article dans une revue
Romanian Journal of Human - Computer Interaction, 2010, Proceedings of the National Conference on Human-Computer Interaction Bucharest, 2-3 September 2010, 3, pp.95-100. 〈http://rochi.utcluj.ro/rrioc/en/index.html〉
Liste complète des métadonnées

Littérature citée [8 références]  Voir  Masquer  Télécharger

https://hal.archives-ouvertes.fr/hal-01079347
Contributeur : Philippe Dessus <>
Soumis le : samedi 1 novembre 2014 - 09:31:22
Dernière modification le : jeudi 21 juin 2018 - 15:28:14
Document(s) archivé(s) le : vendredi 14 avril 2017 - 14:04:36

Fichier

RoCHI-2010-Dascalu.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01079347, version 1

Collections

TICE | LSE | UGA

Citation

Mihai Dascălu, Ștefan Trăușan-Matu, Philippe Dessus. Evaluarea şi sumarizarea automată a conversaţiilor chat. Romanian Journal of Human - Computer Interaction, 2010, Proceedings of the National Conference on Human-Computer Interaction Bucharest, 2-3 September 2010, 3, pp.95-100. 〈http://rochi.utcluj.ro/rrioc/en/index.html〉. 〈hal-01079347〉

Partager

Métriques

Consultations de la notice

216

Téléchargements de fichiers

203