DEARNN: A HYBRID DEEP LEARNING APPROACH FOR CYBER BULLYING DETECTION IN SOCIAL MEDIA

Abstract

Cyberbullying (CB) has become increasingly common on social media platforms. With the rapid growth and widespread use of social media by people of all age groups, it has become essential to make these platforms safer and more secure. This paper presents a hybrid deep learning model called DEA-RNN for detecting cyberbullying on the Twitter social media network. The proposed DEA-RNN model combines an Elman type Recurrent Neural Network (RNN) with an optimized Dolphin Echolocation Algorithm (DEA) to fine-tune the RNN parameters and reduce training time. The DEA-RNN model was evaluated using a dataset of 10,000 tweets and its performance was compared with several existing machine learning and deep learning approaches such as Bi-directional Long Short-Term Memory (Bi-LSTM), RNN, Support Vector Machine (SVM), Multinomial Naive Bayes (MNB), and Random Forest (RF). Experimental results show that the DEA-RNN model performs better than the compared methods in detecting cyberbullying on Twitter. The proposed model demonstrated strong performance, especially in scenario 3, where it achieved an average accuracy of 90.45%, precision of 89.52%, recall of 88.98%, F1-score of 89.25%, and specificity of 90.94%. These results indicate that the DEA-RNN model is an effective and reliable approach for identifying cyberbullying content on social media platforms.