A Deep Learning Approach for Sarcasm Detection in User generated Content

Using sarcasm in social media is a common way to express negative opinions using positive language, making identifying sarcasm an essential part of the sentimental analysis.
Identifying sarcasm is approached as a two-class classification problem (Binary). Both deep learning models and traditional models have been developed using features such as lexical,
semantic, and pragmatic elements. However, sarcasm can be challenging to detect in natural language processing as it involves language usage that is not always straightforward. Despite this, detecting sarcasm can be valuable in many contexts , which includes social media tracking or monitoring, sentimental analysis, and customer support. This research paper proposes a novel approach, BILSTM-GRU architecture, which uses text representations to learn difficult patterns and semantic structures in the text for identifying the sarcastic data. The approach which is going to propose has the ability to improve the accuracy of detecting sarcasm which contributing towards sentiment analysis on social media platform