LABEL-LEX

A number of researchers and students have been involved in the construction of LABEL-LEX, under the direction of Samuel Eleutério and Elisabete Ranchhod. Major contributions have been given by: Jorge Baptista, Helena Freire, Cristina Mota, Paula Carvalho, Pedro Cordeiro and Ana Cristina Oliveira.

Currently, the LABEL-LEX is structured in several inter-related modules, each organized according to the formal complexity of the linguistic objects it contains. The three main modules are:

• LABEL-LEX-sw, which comprises more than 1,500,000 inflected word forms, automatically generated from a lexicon of about 120,000 lemmas; (download)
• LABEL-LEX-mw, which contains about 80,000 multiword lexical units, mainly multiword nouns (76,400), adverbs and adjectives (3,500).
• LABEL-LEX-gr, which includes formal grammars, represented in FSTs. Grammars are used in combination with the dictionaries: a) to resolve ambiguities; b) to analyse and tag different types of linguistic structures.

LabEL's language resources are integrated into three public corpus processors based on automata technology:

• LABEL-T,which is being developed by S. Eleutério at LabEL
• UNITEX, developed by S. Paumier at the Université Marne-laVallée
• INTEX, developed by M. Silberztein at the Université of Franche Comté

Lexical resources are used by EELO, a public service for Portuguese text annotation (tagging) via Internet, developed by the LabEL's team.

LABEL-LEX is also distributed by ELRA, the European Language Resources Association (http://www.elra.fr/).