Comparative Study for Text Chunking Using Deep Learning: Case of Modern Standard Arabic

Abstract

The task of chunking involves dividing a sentence into smaller phrases by identifying a limited amount of syntactic information. This process involves grouping together consecutive words to form phrases, also known as shallow parsing. Chunking does not provide information on the relationships between these phrases. This paper describes our approach to building chunking models for Arabic text using deep learning techniques. We evaluated several training models and compared their results using a rich data set. The results we obtained were highly encouraging when compared to previous related studies.