Detecting Bullying Tweets with LSTM and NLP: A Deep Learning Approach

Bullying on social media platforms is a growing concern due to its serious consequences for victims. However, detecting bullying content on social media platforms is challenging for researchers due to the large volume of data that needs to be manually identified. To address this issue, researchers have developed various techniques using machine learning and Natural Language Processing (NLP) to automatically detect bullying content on social media platforms. In this study, we proposed a deep learning approach using Long Short-Term Memory (LSTM) and NLP to identify bullying tweets on Twitter. The proposed approach involves several stages of processing, including data preprocessing to remove noise and irrelevant information from tweets, feature extraction to capture relevant information such as sentiment, language style, and content of the tweet, and training the LSTM model to classify tweets as bullying or non-bullying. The approach was evaluated on a dataset of 11,923 tweets, with 80% used for training and 20% for testing. Results showed that the approach achieved an accuracy of 94%, with a precision of 93% and recall of 94%. we suggest that LSTM and NLP can be effective tools for identifying bullying content on social media platforms, and their approach can help create a safer and more inclusive online environment by allowing moderators to quickly remove harmful content and take necessary actions to prevent further instances of cyberbullying.