Empirical Perspective on Significant Technological Barriers in Detecting Multilingual Cyberbullying in Nigerian Languages
Domain:
natural language processingRecord type:
Abstract: In recent years, the potential for cyberbullying has grown in lockstep with the growth of social media. Cyberbullying has increased in Nigeria, where 49.6% (104.4 million) of Nigerians have online access and 26.14 percent of African internet users, predominantly Nigerians, have internet access. Cyberbullying occurs in many languages, not just English. Nigeria, on the other hand, has over 525 native languages, making it one of the world's most linguistically varied countries. Cyberbullying thrives in Nigeria due to the large audience, hence it is vital to identify cyberbullying in multiple languages. However, few researchers addressed this puzzle, hence this paper studied the technological barriers to multilingual cyberbullying detection in Nigerian local languages. An empirical perspective of descriptive survey was used to elicit expert-based responses from 60 purposively sampled cybersecurity engineers with 5 years and above expertise in machine learning and cyberbullying detection across five geopolitical zones in Nigeria to actualize the study's purpose. For variable dimension reduction, Principal Component Analysis (PCA) was used, and linear multiple regression was used to predict the model. The study's findings clearly show that the primary technological variables impeding cyberbullying detection in both English and any of Nigeria's primary indigenous languages of Hausa, Yoruba, and Igbo are language ambiguity, sarcasm dialect classification, domain influence, data scarcity, non-trusted source of data, difficulty in developing novel classifiers for multilingual text detection, annotation of training data, and language characteristics. It is recommended that study be performed to discover latent cyberbullying actions that cannot be identified in English Languages.
Keywords: Barriers, Cyberbullying, Detection, Multilingual, Nigerian Languages, Technology.
Title: Empirical Perspective on Significant Technological Barriers in Detecting Multilingual Cyberbullying in Nigerian Languages
Author: Christiana A. Okoloegbo, Udoka F. Eze, Gloria A. Chukwudebe,, Obi C. Nwokonkwo
International Journal of Novel Research in Computer Science and Software Engineering
ISSN 2394-7314
Vol. 10, Issue 3, September 2023 - December 2023
Page No: 7-15
Novelty Journals
Website: www.noveltyjournals.com
Published Date: 16-October-2023
DOI: https://doi.org/10.5281/zen…
Paper Download Link (Source)