Analysis and Mining of Arabic Comparative Sentences: A Literature Review

Document Type : Original Article

Authors

Computer Science, Faculty of Computers and Information, Menoufia University

Abstract

There are huge Arabic comparative sentences that are generated daily on various social media. These comparative sentences need to be analyzed and mined for different purposes such as service and product reviews. In general, analysis and mining of Arabic text is a big challenge due to the limitations inherited in the Arabic language. Moreover, there are currently no standard datasets for Arabic comparative sentences. This paper provides a background on the different steps required to analyze and mine an Arabic comparative sentence. These steps include the identification of the Arabic comparative sentence, the identification of the sentence type, and finally the extraction of a relation together with a preferred entity. The paper also provides a literature review of current research work applied in this research field. This includes a classification of the various techniques leveraged in this field including three main categories namely: linguistic, machine learning and deep learning approaches. Finally, the paper provides insights on current limitations and future research challenges in this field. To the best of our knowledge, this is the first research paper that provides a dedicated literature review about the analysis and mining of Arabic comparative sentences. This review discusses the specific analysis of Arabic comparative sentences not the general Arabic sentiment analysis. It is noted that this analysis is a subset of the Arabic sentiment analysis field which does not focus on identifying the sentiment of an Arabic sentence, however, it focuses on identifying and analyzing an Arabic comparative sentence and its components.

Keywords

Files

Share

How to cite

Statistics