Corpus-based Methodology for an Online Multilingual Collocations Dictionary: First Steps

• Autor(a) principal: Orenha-Ottaiano, Adriane [UNESP]
• Data de Publicação: 2021
• Outros Autores: Garcia, Marcos, de Oliveira Silva, Maria Eugênia Olímpio, L'Homme, Marie-Claude, Ramos, Margarita Alonso, Valêncio, Carlos Roberto [UNESP], Tenório, William [UNESP]
• Tipo de documento: Artigo de conferência
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/242228
• topic: automatic extraction; collocations; collocations dictionary; lexicography; online platform

This paper describes the first steps of a corpus-based methodology for the development of an online Platform for Multilingual Collocations Dictionaries (PLATCOL). The platform is aimed to be customized for different target audiences according to their needs. It covers various syntactic structures of collocations that fit into the following taxonomy: verbal, adjectival, nominal, and adverbial. Part of its design, layout and methodological procedures are based on the Bilingual Online Collocations Dictionary Platform (Orenha-Ottaiano, 2017). The methodology also relies on the combination of automatic methods to extract candidate collocations (Garcia et al., 2019a) with careful post-editing performed by lexicographers. The automatic approaches take advantage of NLP tools to annotate large corpora with lemmas, PoS-tags and dependency relations in five languages (English, French, Portuguese, Spanish and Chinese). Using these data, we apply statistical measures (Evert et al., 2017; Garcia et al., 2019b) and distributional semantics strategies to select the candidates (Garcia et al., 2019c) and retrieve corpus-based examples (Kilgarriff et al., 2008). We also rely on automatic definition extraction (Bond & Foster, 2013) so that collocations can be more effectively organized according to their specific senses.