Reason Based Machine Learning Approach to Detect Bangla Abusive Social Media CommentsShow others and affiliations
2023 (English)In: Intelligent Computing & Optimization: Proceedings of the 5th International Conference on Intelligent Computing and Optimization 2022 (ICO2022) / [ed] Pandian Vasant; Gerhard-Wilhelm Weber; José Antonio Marmolejo-Saucedo; Elias Munapo; J. Joshua Thomas, Springer, 2023, 1, p. 489-498Chapter in book (Refereed)
Abstract [en]
For the study issue of abusive language detection, English is the most commonly employed language. There are just a few works accessible in low-resource languages such as Bangla. People use these sorts of statements on many social media sites. As a result, detection of this type of language is a demand of time. Our goal is to identify this abusive Bangla language in a novel approach. There are some works that use Bengali corpus and transliterated Bengali corpus to detect abusive language. However, in this research, we utilized annotated translated Bengali corpora, and we added a formal justification in each remark for being classified as abusive or non abusive language. For evaluations, we employed a variety of machine learning classifiers where logistic regression achieves 97% accuracy.
Place, publisher, year, edition, pages
Springer, 2023, 1. p. 489-498
Series
Lecture Notes in Networks and Systems (LNNS), ISSN 2367-3370, E-ISSN 2367-3389 ; 569
Keywords [en]
Low-resource language, Machine learning, Social media comments
National Category
Computer Sciences
Research subject
Pervasive Mobile Computing
Identifiers
URN: urn:nbn:se:ltu:diva-94215DOI: 10.1007/978-3-031-19958-5_46Scopus ID: 2-s2.0-85144511050OAI: oai:DiVA.org:ltu-94215DiVA, id: diva2:1712547
Note
ISBN för värdpublikation: 978-3-031-19958-5; 978-3-031-19957-8
2022-11-222022-11-222024-03-07Bibliographically approved