MS or Mphil

Permanent URI for this community

Now showing 1 - 1 of 1

A CYBERBULLYING MULTICLASS CLASSIFICATION WITH AN ATTENTION MECHANISM HANDCRAFTED CNNBERT WITH BILSTM
(UMT, Lahore, 2023) Ammara Asghar
Cyberbullying is an important problem that affects people across the globe and is unpleasant for both the individuals who are bullied and society overall. One type of bullying that occurs on several social networking platforms is referred to as cyberbullying. Twitter has become one of the most well-known locales for posting recordings, with a great many tweets. AI and normal language handling procedures are utilized to investigate the remarks. However, for cyberbullying analysis, many machine learning models analyze comments in fixed-length frames or sequences, which results in the loss of long-term context information. This postulation explores an original technique for recognizing cyberbullying that consolidates Convolutional Brain Organizations with Consideration System BERT (CNNBERT) and Bi-directional Long Momentary Memory (BiLSTM) models. This exploration utilizes a half and half profound learning way to deal with create a complex model that can precisely distinguish and characterize cyberbullying content across online entertainment stages. The proposed technique performs effectively by evaluating text-based data to identify complex instances of cyberbullying. Combining BiLSTM expert ability to process sequential input with the attention mechanism of BERT and CNN, which enhances feature extraction.