Implementing A Scheme That Uses Recurrent Neural Networks (RNNS) And Linear Support Vector Machines (LSVM) To Uncovers Online Bullying on Social Media Platforms

Abstract

In recent years, one of the main challenges in establishing a safe online environment has been the development of a plan to identify and handle online bullying on social media platforms. The plan outlines a study that used a combination of recurrent neural networks (RNNs) and linear support vector machines (LSVM) to create a cyberbullying detection system. The idea was to develop a system that could discover and identify people who were bullying others online. Using a dataset of text expressions from Twitter and Facebook, the study's algorithm successfully classified instances of cyberbullying; the LSVM achieved an 85% classification accuracy, while the RNN achieved an 81% feature extraction accuracy. This performed better than both deep learning and conventional machine learning techniques. The LSVM and RNN algorithms, according to the authors, have the ability to effectively handle the issue of online cyberbullying. They make a number of suggestions to improve the system even further, including developing more user-friendly interfaces, enhancing dataset diversity, giving data protection first priority, and upgrading the model on a regular basis. The authors also recommend several avenues for further research, such as combining multimodal approaches, putting real-time monitoring and intervention into practice, analyzing data across platforms, and assessing the system's long-term effects. All things considered, the study shows how promising it is to use cutting-edge AI methods like LSVM and RNNs to create practical answers to the urgent problem of cyberbullying.