ALIF editor for generating Arabic normalized lexicons

Abstract

The development of a normalized morpho-syntactic Arabic lexicon is not an easy task. In fact, many norms allow the structuration and representation of lexical data. The adoption of a stable standard will guarantee the interoperability and interchangeability of lexical resources. Still, research work that deals with normalization for Arabic lexical resources is not well developed yet, especially for some standards such as the TEI (Text Encoding Initiative). In this context, we aim at creating an Arabic lexicon editor with a constraint checker based on both the ISO standard LMF (Lexical Markup Framework) and the TEI guidelines. To develop this editor, we use a linguistic approach composed of several steps. The editor's prototype named ALIF can guarantee the construction of two types of output lexicon files: one in LMF and the other in TEI. The evaluation of this system is based upon a lexical database that contains all the derived and inflected forms generated from a lexicon of 10 000 canonical verbs. The results obtained were encouraging despite some flaws related to exceptional cases of difficult words.