Researching the Arabic dialects, including the Algerian dialect, and studying the texts written in them constitutes a bet for natural language processing and carrying out the tasks needed by researchers and users, especially after the spread of posts about them on social networks. As a
result, problems appeared that impede the process of recognizing written words and phrases in these dialects, including the problem of ambiguity in the absence of corpus, problems with orthographic systems, and the overlap between them and standard Arabic or foreign languages.
Accordingly, I studied a sample of the postings of students of the Department of Arabic Language and Literature at Mascara University, written in Darija and published on their Facebook page, and noting the difficulties they show in the process of recognizing their words, which prompted researchers to adopt deep learning to achieve a percentage of success.

---