Intelligent Systems for Cyberbullying Detection and Response

Recently, social media and online communication tools

have become an integral part of digital technology, and the

widespread use of the same has definitely increased the rate of

digital interactions. However, it has also increased the rate of cases

of cyberbullying and other types of abusive online

communications. Online communications, such as abusive words,

harassment, and also the use of slangs and other sorts of abusive

words that reflect upon the psychological state of individuals, are

on the rise and are having a major impact on individuals.

Traditional mechanisms of identifying online abusive

communications are either not effective enough to work efficiently

with a large database of communications or are not aware of the

contexts. In this regard, this paper proposes a framework of a web

based cyberbullying detection tool known as CyberGuardian that

attempts to identify abusive communications submitted to the

system by the users. In this paper, we proposed the development of

a tool that aids in checking user-submitted communications to

identify abusive words and context. The tool has been implemented

with a modular Flask framework and can be said to have low

computational complexities, and the framework has been highly

effective in identifying explicit abusive words.

Index terms: Cyberbullying Detection, Toxic Language Analysis,

Natural Language Processing, Text Classification, Web Application,

Online Safety.