A Hybrid Classifier of Cyber Bullying Detection in Social Media Platforms

Abstract

The increase in online and social media connection has made it simple for hate speech and insulting language to spread. Cyberbullying is the phrase used to describe such online abuse, insults, and assaults. It has become difficult to detect such unauthorized material due to the huge number of user-generated content. Deep neural networks are being used more often by academics to identify cyberbullying than regular machine learning algorithms because to their many benefits over them. Machine learning has several uses in text categorization. Hence, it is fundamental to distinguish and sort CB utilizing profound learning (DL) models in informal organizations to avoid this pattern. FSSDL-CBDC is a fresh out of the plastic new methodology for informal communities that joins profound learning and element subset determination. The SSA-DBN model has demonstrated to be more exact than different calculations with a 99.983% precision rate. Generally speaking, the trials' discoveries showed that the FSSDL-CBDC strategy performs better compared to the contending systems in various ways.