Cyberbullying-Related Hate Speech Fine-Grained Classification for Social Media Forensics Using Neutrosophic Neural Networks

Detecting cyberbullying automatically in social media has emerged as a highly challenging task due to the complex nature of language used in such platforms. Currently, several methods exist for cyberbullying detection, but they still suffer from the ambiguity and vagueness when differentiating between different types of cyberbullying-related hate speech and they also lack in accuracy. This paper proposes a fine-grained cyberbullying classification approach by integrating Neutrosophic Logic (NL) within the Multi-Layer Perceptron (MLP) model. The proposed model enhances classification by mitigating the challenges posed by the ambiguity and overlapping boundaries between distinct categories of cyberbullying. The model, leveraging the one-against-one strategy in MLP classification, captures complex relationships between various types of cyberbullying, due to the overlaps and ambiguous instances within cyberbullying types. The results of the proposed model demonstrate the performance enhancement of incorporating Neutrosophic Logic for fine-grained cyberbullying classification tasks, with a 2% increase in accuracy, thereby improving the overall effectiveness of cyberbullying classification.