Scopus Indexed Publications

Paper Details


Title
Bengali Slang detection using state-of-the-art supervised models from a given text
Author
Md. Abdul Hamid, Atiqur Rahman, Eteka Sultana Tumpa, Jabir Al Nahian, Johora Akter Polin, Nurjahan Akther Mim,
Email
ahmadhamid040@gmail.com
Abstract

Almost all Bengalis who own smartphones also have social media accounts. People from different regions occasionally employ regional Slang that is unfamiliar to outsiders and confuses the meaning of the sentence. Nearly all languages can now be translated thanks to modern technology, but only in very basic ways, which is a concern. Bengali Slang terms are difficult to translate due to a dearth of rich corpora and frequently occurring new Slang terms developed by people, making it impossible for speakers of other languages to understand the context of a sentence in which Slang is used. We developed a solution to this issue. To create models that can detect Bengali Slang terms from social media, we gather various Slang phrases from various regions and develop a modest corpus. Our suggested method nearly always succeeds in extracting Bengali Slang terms from fresh material. We create a total of 7 supervised models and assess which is the most effective for our study. One of them has a 70% accuracy and 86% recall rate for successful identification. Our models may be linked to the social media platform's backend to restrict the use of Bengali Slang in posts, blogs, comments, and other areas.

Keywords
Journal or Conference Name
Bulletin of Electrical Engineering and Informatics
Publication Year
2023
Indexing
scopus