Automatic Dialog Acts Recognition based on Words Clusters - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2006

Automatic Dialog Acts Recognition based on Words Clusters

Pavel Kral
  • Fonction : Auteur
  • PersonId : 831051
Jana Kleckova
  • Fonction : Auteur
  • PersonId : 831052

Résumé

This paper deals with automatic dialog acts (DAs) recognition in Czech. A Dialog act is defined by J. L. Austin [1] as a meaning of an utterance at the level of illocutionary force. The four following DAs are considered: statements, orders, yes/no questions and other questions. In our previous works, we proposed, implemented and evaluated two new approaches to automatic DAs recognition based on sentence structure. These methods have been validated on a Czech corpus that simulates a task of train tickets reservation. The main goal of this paper is to propose a new approach to solve the problem of lack of training data for automatic DA recognition. This approach clusters the words in the sentence into several groups using maximization of mutual information between two neighbor word classes. The classification accuracy of the unigram model (our baseline approach) is 91 %. The proposed method, a clustered unigram model, reduces the DA error rate by 12 %
Fichier principal
Vignette du fichier
wespacIX.pdf (189.71 Ko) Télécharger le fichier
Loading...

Dates et versions

hal-00086310 , version 1 (18-07-2006)

Identifiants

  • HAL Id : hal-00086310 , version 1

Citer

Pavel Kral, Jana Kleckova, Christophe Cerisara. Automatic Dialog Acts Recognition based on Words Clusters. 2006, 6 p. ⟨hal-00086310⟩
183 Consultations
130 Téléchargements

Partager

Gmail Facebook X LinkedIn More