Challenges in building corpora for Algerian Arabic from CMC content

Abstract EN

Algerian Arabic is an under-resourced Arabic dialect.

few corpora and natural language processing tools were developed for it.

this is due to a variety of factors such as its lack of written content and of a standard orthography as well as the frequent code-switching and script switching exhibited by its speakers.

these factors render developing homogenous corpora for the dialect more challenging compared to other Arabic dialects where such factors are less pronounced.

the objective of this work is to examine the challenges and issues encountered in developing a corpus of Algerian Arabic extracted from computer-mediated communication content, primarily content on the social media platform Facebook and the story-publishing website Wattpad.