Generating a Lexicon for the Hijazi Dialect in Arabic

Abstract

We present a methodology for creating a lexicon for a low-resource Arabic dialect in Saudi Arabia: Hijazi. We show the differences between the Hijazi dialect and Modern Standard Arabic. We annotate articles and tweets using recruited native speakers. We create a lexicon of Hijazi adapted from two resources: Sebawai and Quranic Arabic Corpus. The lexicon is created both manually and automatically by using Hijazi morphology. We detail the methodology to build this lexicon and present results of an evaluation of the corpus formation process.