SentiALG: Automated Corpus Annotation for Algerian Sentiment Analysis

Guellil, Imane; Adeel, Ahsan; Azouaou, Faical; Hussain, Amir

Abstract

Data annotation is an important but time-consuming and costly procedure. To sort a text into two classes, the very first thing we need is a good annotation guideline, establishing what is required to qualify for each class. In the literature, the difficulties associated with an appropriate data annotation has been underestimated. In this paper, we present a novel approach to automatically construct an annotated sentiment corpus for Algerian dialect (A Maghrebi Arabic dialect). The construction of this corpus is based on an Algerian sentiment lexicon that is also constructed automatically. The presented work deals with the two widely used scripts on Arabic social media: Arabic and Arabizi. The proposed approach automatically constructs a sentiment corpus containing 8000 messages (where 4000 are dedicated to Arabic and 4000 to Arabizi). The achieved F1-score is up to 72% and 78% for an Arabic and Arabizi test sets, respectively. Ongoing work is aimed at integrating transliteration process for Arabizi messages to further improve the obtained results.

• Presentation Conference Type: Conference Paper (Published)
• Conference Name: BICS: International Conference on Brain Inspired Cognitive Systems
• Start Date: Jul 7, 2018
• End Date: Jul 8, 2018
• Online Publication Date: Oct 6, 2018
• Publication Date: 2018
• Deposit Date: Jul 26, 2019
• Journal: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
• Publisher: Springer
• Pages: 557-567
• Series Title: Lecture Notes in Computer Science
• Series Number: 10989
• Series ISSN: 0302-9743
• ISBN: 978-3-030-00562-7
• DOI: https://doi.org/10.1007/978-3-030-00563-4_54
• Keywords: Arabic sentiment analysis, Algerian dialect, Sentiment lexicon, Sentiment corpus, Sentiment classification
• Public URL: http://researchrepository.napier.ac.uk/Output/1792340