Hyperbolic Feature-based Sarcasm Detection in Tweets: A Machine Learning Approach

Abstract

In recent times, sarcasm analysis has been one of the toughest challenges in Natural Language Processing (NLP). The property of sarcasm that makes it difficult to analyze and detect is the gap between its literal and intended meaning. Detecting sarcastic sentiment in the domain of social media such as Facebook, Twitter, online blogs, reviews, etc. has become an essential task as they influence every business organization. In this article, a hyperbolic feature-based sarcasm detector for Twitter data is proposed. The hyperbolic features consist of intensifiers and interjections of the text. The performance of the proposed system is analyzed using several standard machine learning approaches namely, Naive Bayes (NB), Decision Tree (DT), Support Vector Machine (SVM), Random Forest (RF), and AdaBoost. The system attains an accuracy (%) of 75.12, 80.27, 80.67, 80.79, and 80.07 using NB, DT, SVM, RF, and AdaBoost respectively.