Prediction of Cyberbullying Attacks on Twitter Data using ANN and NLP
Israel Nixon, Deva Priya Isravel, Julia Punitha Malar Dhas · 2024
As cyberbullying has become more common, people are becoming more worried about the damage it may do to individuals and communities. This work aims to develop a model for predictive analysis that can scour Twitter data for patterns and foretell occurrences of cyberbullying. Therefore, a novel framework with a hybrid of ANN and NLP model is proposed to assess Twitter data. The proposed hybrid ANN-NLP model is fine-tuned by using the Vectorizer algorithm to detect and classify instances of cyberbullying more accurately. The efficiency of the proposed model is experimented on a dataset that has abusive or otherwise inappropriate language. The model exhibited better performance metrics in terms of precision, recall, f1-score and accuracy. The hybrid ANN-NLP model obtained an accuracy of 91% when compared with other classical methods. The results provide light on the features and patterns of cyberbullying and are useful for detecting crimes.