Home |  English version |  Mappa |  Commenti |  Sondaggio |  Staff |  Contattaci Cerca nel sito  
Istituto di linguistica computazionale "Antonio Zampolli"

Torna all'elenco Contributi in atti di convegno anno 2005

Contributo in atti di convegno

Tipo: Contributo in atti di convegno

Titolo: Automatic Incremental Term Acquisition from Domain Corpora

Anno di pubblicazione: 2005

Formato: Cartaceo

Autori: Bartolini R., Giorgetti D., Lenci A., Montemagni S., Pirrelli V.

Affiliazioni autori: Lenci A. (Università di Pisa).

Autori CNR:


Lingua: inglese

Abstract: We describe a technique for the acquisition of terms from Italian domain text corpora, which relies both on sophisticated linguistic analysis and on statistical measures applied to linguistically processed text rather than to raw text as it is usually the case. The main advantage of this technique is that minimal a priori knowledge of term structure is required, thus allowing to explore and discover terms in a given domain without imposing a strict pattern matching structure on them, and also to easily extend it to different domains. The approach we present in this paper is incremental as it may be iterated to discover terms of increasing complexity built on top of terms discovered in the previous iteration. The reason why it is convenient to adopt such an incremental approach is that it allows to "clean" data from noise in the first step, elicitating the constituent terms, and then to refine term acquisition on "skimmed" term data.

Lingua abstract: inglese

Pagine da: 293

Pagine a: 300

Titolo del volume: Proceedings of TKE 2005 - 7th International Conference on Terminology and Knowledge Engineering

Referee: Sì: Internazionale

Congresso nome: 7th International conference on Terminology and Knowledge Engineering (TKE2005)

Congresso luogo: Copenhagen

Congresso rilevanza: Internazionale

Congresso relazione: Contributo

Strutture CNR:


Torna indietro Richiedi modifiche Invia per email Stampa
Home Il CNR  |  I servizi News |   Eventi | Istituti |  Focus