Detection of Cyberbullying Using Modified Dense Framework
Varun Malik, Ruchi R. Mittal, Vikram Sıngh, Amit Mittal, S Vikram Singh, Shashi Prakash Diwvedi · 2023
Cyberbullying has emerged as a significant societal concern, with the proliferation of digital communication platforms. Detecting and mitigating cyberbullying is crucial to ensure online safety and promote positive digital interactions. This study proposes an innovative approach for the detection of cyberbullying using a Modified Dense Framework (MDF). The Modified Dense Framework (MDF) introduced in this research is a deep learning architecture specifically tailored for cyberbullying detection. Unlike traditional methods, MDF combines the power of convolutional neural networks (CNNs) and recurrent neural networks (RNNs) in a novel way, enhancing the model's ability to capture intricate patterns within textual and multimedia content. The framework is trained on diverse datasets encompassing text, images, and videos, enabling it to effectively identify various forms of cyberbullying across multiple online platforms. The study demonstrates the effectiveness of MDF through rigorous experimentation and evaluation. Comparative analyses with existing state-of-the-art models highlight the superiority of MDF in terms of accuracy, precision, recall, and F1-score. Moreover, MDF exhibits remarkable resilience against adversarial attacks, ensuring robust cyberbullying detection even in challenging scenarios.