Sarcasm Detection of Textual Data on Online SocialMedia: A Review

This paper refers to the sarcasm detection process and approaches and comparison of results on various models and datasets. Sarcasm refers to the phrases which indicates opposite meaning of what it actually wants to express. In the recent years the NLP became the very interesting topic for the researchers. Sarcasm detection is also part of NLP. Sarcasm detection is somewhat similar to sentiment analysis which mathematically illustrates and categorize the polarity of piece of text or phrase and determine whether it is sarcastic or not. In recent years sarcasm detection was performed on twitter datasets reddit corpus, SARC dataset and many more. Main focus of this paper is on various ml and deep learning approaches to sarcasm detection like Support Vector Machine, Convolution Neural network, LSTM models used for sarcasm detection in recent researches. In our study we will explore various approaches to sarcasm detection. At the end I will compare and contrast the different approaches for sarcasm detection based on their accuracy.