Bridging NLP Gaps: Methodology to Create A Novel Dataset for Named Entity Recognition in Algerian Dialectal Arabic

Abstract

This paper introduces a methodology to create a dataset designed to address the challenges of Named Entity Recognition (NER) in Algerian Dialectal Arabic (ADA). While significant advancements have been made in NER for Modern Standard Arabic (MSA), dialectal varieties like ADA remain largely underrepresented in natural language processing (NLP) resources. To bridge this gap, we propose a systematic method for collecting and annotating ADA texts. The dataset will be curated from informal sources, including social media platforms, where ADA is predominantly used