Arabic text classification methods: Systematic literature review of primary studies

W. Alabbas, Haider Al-Khateeb, A. Mansour

Research output: Chapter in Book/Published conference outputConference publication

Abstract

Recent research on Big Data proposed and evaluated a number of advanced techniques to gain meaningful information from the complex and large volume of data available on the World Wide Web. To achieve accurate text analysis, a process is usually initiated with a Text Classification (TC) method. Reviewing the very recent literature in this area shows that most studies are focused on English (and other scripts) while attempts on classifying Arabic texts remain relatively very limited. Hence, we intend to contribute the first Systematic Literature Review (SLR) utilizing a search protocol strictly to summarize key characteristics of the different TC techniques and methods used to classify Arabic text, this work also aims to identify and share a scientific evidence of the gap in current literature to help suggesting areas for further research. Our SLR explicitly investigates empirical evidence as a decision factor to include studies, then conclude which classifier produced more accurate results. Further, our findings identify the lack of standardized corpuses for Arabic text; authors compile their own, and most of the work is focused on Modern Arabic with very little done on Colloquial Arabic despite its wide use in Social Media Networks such as Twitter. In total, 1464 papers were surveyed from which 48 primary studies were included and analyzed.
Original languageEnglish
Title of host publicationColloquium in Information Science and Technology, CIST
Place of PublicationTangier, Morocco
PublisherIEEE
Pages361 - 367
Number of pages6
ISBN (Electronic)978-1-5090-0751-6
ISBN (Print)978-1-5090-0752-3
DOIs
Publication statusPublished - 2017
Event2016 4th IEEE International Colloquium on Information Science and Technology (CiSt) - Tangier, Morocco, Tangier, Morocco
Duration: 24 Oct 201626 Oct 2016

Conference

Conference2016 4th IEEE International Colloquium on Information Science and Technology (CiSt)
Country/TerritoryMorocco
CityTangier
Period24/10/1626/10/16

Keywords

  • Arabic text classification
  • big data
  • systematic literature review
  • data mining
  • Text corpus

Fingerprint

Dive into the research topics of 'Arabic text classification methods: Systematic literature review of primary studies'. Together they form a unique fingerprint.

Cite this