Comparative Study on Performance of Machine Learning Algorithm and Deep Learning Algorithms for Sarcasm Detection

Abstract

Sarcasm is irony, to mock or deliver contempt. It is delivered in comic and sometimes in terrible tone which is often not taken in positive sense. This sarcasm has become a trend to mock or discourage people. Hence it is necessary to detect sarcasm in text which is mostly observed in social media platforms like Facebook, twitter, Instagram etc. Machine learning has become an aid in analyzing textual information for sentiment analysis and sarcasm detection. Recent research mainly focuses on deep learning approaches due to its promising results. This paper presents a comparison of various machine learning and deep learning approaches relevant to sarcasm detection. Further, results of these approaches and their evaluation are presented and future directions are mentioned. In this paper, we have used five methods of machine learning to detect sarcasm. Deep learning methods are also being used with machine learning methods and compared.