PUMA
Istituto di Informatica e Telematica     
Del Gratta R., Frontini F., Monachini M., Quochi V., Rubino F., Abrate M., Lo Duca A. L-LEME: an Automatic Lexical Merger based on the LMF Standard. In: LREC 2012 - Language Resources and Evaluation (Istanbul, 2012). Proceedings, pp. 31 - 40. Núria Bel - Universitat Pompeu Fabra, Barcelona, Spain; Maria Gavrilidou - ILSP/Athena R.C., Athens, Greece; Monica Monachini - CNR-ILC, Pisa, Italy; Valeria Quochi - CNR-ILC, Pisa, Italy; Laura Rimell - University of Cambridge, UK, 2012.
 
 
Abstract
(English)
The present paper describes LMF LExical MErger (L-LEME), an architecture to combine two lexicons in order to obtain new resource(s). L-LEME relies on standards, thus exploiting the benefits of the ISO Lexical Markup Framework (LMF) to ensure interoperability. L-LEME is meant to be dynamic and heavily adaptable: it allows the users to configure it to meet their specific needs. The L-LEME architecture is composed of two main modules: the Mapper, which takes in input two lexicons A and B and a set of user-defined rules and instructions to guide the mapping process (Directives D) and gives in output all matching entries. The algorithm also calculates a cosine similarity score. The Builder takes in input the previous results, a set of Directives D1 and produces a new LMF lexicon C. The Directives allow the user to define its own building rules and different merging scenarios. L-LEME is applied to a specific concrete task within the PANACEA project, namely the merging of two Italian SubCategorization Frame (SCF) lexicons. The experiment is interesting in that A and B have different philosophies behind, being A built by human introspection and B automatically extracted. Ultimately, L-LEME has interesting repercussions in many language technology applications.
Subject Language Technologies
Lexicon Merging
LMF Standard
Similarity Score
H.3.1 Linguistic processing


Icona documento 1) Download Document PDF


Icona documento Open access Icona documento Restricted Icona documento Private

 


Per ulteriori informazioni, contattare: Librarian http://puma.isti.cnr.it

Valid HTML 4.0 Transitional