Twitter Sarcasm Detection using Natural Language Processing and Deep Learning Techniques

In the last ten years, sarcasm detection has been an increasingly popular subject in natural language processing (NLP), which studies how language is used to convey meaning. Promising results have been achieved by a wide range of methods, from conventional computer learning to multifaceted deep learning techniques. Many methods have arisen that rely on context to determine whether or not a given piece of text contains sarcasm. However, sarcasm-context detection has not been the primary focus of any natural language processing studies. This research provides a method for automated sarcasm context detection with the goal of creating models that can accurately determine whether sarcasm is present or suitable. The Weighted Random Forest approach is used to train and compare many models for sarcasm-context detection utilising a widely used dataset (MUStARD). This top-notch performance is established as an F1 score of 60,1 achieving attention-based short-term and long- architecture. In addition, we evaluated the model's efficacy by testing it on the SARC database & contrasting the outcomes with those previously published in the literature. The possibility of creating a conversational AI that can recognise and react to sarcasm opens up new avenues of research.