Sarcasm Detection Using Machine Learning

Abstract

Now a day’s sarcasm detection has been recognized as one of the toughest challenge of Natural
Language Processing (NLP). The sarcasm detection becomes tough due to the gap between what is said and
what it really means.In the field of social media such as Facebook, Whatsapp, Instagram, Twitter and
reviews, it has became very important to detect the sarcastic comments as it effect the ratings of business
activities for example ratings of amazon products. In our research we are proposing a sarcasm detector to
analyze the variety of comments from different platforms such as:- Twitter, Facebook, Whatsapp, and selfgenerated records. The proposed model is tested for different machine learning techniques namely, Support
Vector Machine (SVM), Convolutional Neural Network (CNN) and comprehensive performance is
analyzed. The system acquires an accuracy (%) of 99%,97%, the use of SVM,CNN respectively