Analysis on Text for Sarcasm Detection Using Machine Learning: Survey

Abstract

Nowadays, people using sarcasm in their daily communication with others. In person communication sarcasm detection is not that complex but text communication it’s very difficult. People share their views, opinions, comments on others post over social media like Twitter, Reddit, Facebook, YouTube, LinkedIn, Media, Blogs, Discussion forums and WhatsApp etc. Researcher have been applied so many approaches to detect sarcasm in text but not yet reached 100% accuracy. It’s a challenging task to researchers finding sarcasm on text like social media platforms, newspapers, product reviews, movies reviews, comments on YouTube and feedbacks etc. Sarcasm in text can be irony, humour and criticism with positive words convey negative meaning. Sarcastic statements create confusion whether sarcasm used or not in their statement. So far Research have done on sarcasm detection with text information but emojis, numbers, empty messages also conveying sarcasm.  In this paper, addressing approaches and challenges to find sarcasm detection and performance measurements of existing research like recall, precision and accuracy.