DEARNN A HYBRID DEEP LEARNING APPROACH FOR CYBERBULLYING DETECTION IN TWITTER SOCIAL MEDIA PLATFORM

Abstract

Cyberbullying has emerged as a significant social issue with the rapid growth of social media platforms such as Twitter, where users frequently express opinions and interact in real time. Detecting harmful and abusive content in such dynamic environments is challenging due to the informal language, slang, sarcasm, and contextual variations present in online communication. This project proposes DEARNN (Deep Ensemble Attention-based Recurrent Neural Network), a hybrid deep learning approach designed to accurately detect cyberbullying in Twitter data. The model combines the strengths of multiple deep learning architectures, including Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), and attention mechanisms, to effectively capture both sequential patterns and contextual semantics in textual data. The proposed system begins with data collection from Twitter datasets, followed by preprocessing steps such as text cleaning, tokenization, stopword removal, and word embedding using techniques like Word2Vec or GloVe. The DEARNN model processes the input text by leveraging sequential learning through LSTM layers while the attention mechanism focuses on critical words or phrases that contribute most to cyberbullying detection. The ensemble component integrates outputs from multiple models to improve robustness and prediction accuracy. This hybrid architecture enables the system to handle complex linguistic patterns, sarcasm, and contextual dependencies more effectively than traditional machine learning models. The performance of the model is evaluated using standard metrics such as accuracy, precision, recall, and F1-score. Experimental results demonstrate that DEARNN outperforms baseline models, achieving higher detection rates and reduced false positives. The system can be deployed in real-time social media monitoring tools to automatically identify and filter harmful content, thereby promoting a safer online environment. This research contributes to the advancement of intelligent content moderation systems and highlights the importance of hybrid deep learning techniques in addressing complex natural language processing challenges.Keywords: Cyberbullying Detection,