Prolmf, a Multilingual Dictionary of Proper Names and their Relations
Résumé
This chapter presents Prolmf, a Lexical Markup Framework (LMF)-conformant multilingual dictionary of proper nouns and their relations. It is part of a French project called Prolex that aims to build an Natural Language Processing (NLP) platform to process proper nouns in a multilingual context. The main resources issued from Prolex are (1) a cascade of transducers to recognize French-Named Entities in texts and (2) a relational database called Prolexbase. Prolmf is the part of Prolexbase that is published on the Web. It relies on LMF in order to enhance data exchange and data integration. The chapter explains the monolingual part and the multilingual part, before focusing on the choices of representation for the three new semantic descriptions recently added: the classifying contexts of proper names, the accessibility relation between two proper names and idiosyncratic collocations (determiners and/or locative prepositions).