Due to the Covid19 pandemic, and the restrictions placed in social interactions, there has been an upsurge in the use of social networks, such as Facebook, WhatsApp, Instagram, Telegram, and Twitter to name a few. As more and more people turn to the social networks for social interaction, there has been increased occurrences of cyberbullying. Cyberbullying is a type of bullying that occurs through online technology, whereby harmful texts and pictures are shared through social networks. This research project aims to develop a system that can detect cyberbullying on social networks such as Twitter focusing on the IsiXhosa language. Machine learning algorithms will be applied to Twitter feeds dataset to detect cyberbullying. The project will help the law enforcement to be able to catch cyberbullies that make threats using isiXhosa.
To Run the Detection System: You need to have a python IDE
- First run the Translating.py file which collects Xhosa tweets from twitter: the code will ask you to enter a word, which is any Xhosa bullying word you know to extract the data. Words such as: Msunu, Mnqundu, etc
- Thereafter, run the Translating Data.py file, which will translate the xhosa data into English so that we can detect
- Run the Clean.ipynb and then run the Cyberbullying detection code
The 2 last codes labeled data extracted from Github which is labeled_data nd plabeled