Multilingual Cyberbullying Detector (CD) Application for Nigerian Pidgin and Igbo Language Corpus

Abstract

In recent years, difficulties related to cyberbullying have emerged as a result of the expansion of social media platforms and community interaction. Naive Bayes classifiers and other well-known models have been used successfully by several academics to create sentiment analysis systems for various use cases. Recent advances in the detection and management of multilingual cyberbullying actions on forums and social networking sites have built on the success of these sentiment analysis efforts. In order to reduce cybercrime in Nigeria, the study's goal is to create an improved Cyberbullying Detector (CD) that is interactive, affordable, and helps identify, monitor, and regulate cyberbullying. The application is the first of its kind in Nigeria to monitor and regulate cyberbullying on Twitter in Pidgin English and Igbo Language. A custom pidgin library was developed with comprehensive translations. The TextBlob library is appropriate for the study, which focuses on cyberbullying, in terms of sentiment prediction. From the sentiment analysis of Twitter data collected using SNScrape, the results show language-specific models that worked perfectly in flagging cyberbullying at manageable runs.