A Pattern-Based Approach for Sarcasm Detection on Twitter

• Published in: IEEE access Vol. 4; pp. 5477 - 5488
• Main Authors: Bouazizi, Mondher, Otsuki Ohtsuki, Tomoaki
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2016; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Data mining; Feature extraction; Machine learning; sarcasm detection; Sentiment analysis; Social networks; Twitter; Websites
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2016.2594194

Sarcasm is a sophisticated form of irony widely used in social networks and microblogging websites. It is usually used to convey implicit information within the message a person transmits. Sarcasm might be used for different purposes, such as criticism or mockery. However, it is hard even for humans to recognize. Therefore, recognizing sarcastic statements can be very useful to improve automatic sentiment analysis of data collected from microblogging websites or social networks. Sentiment Analysis refers to the identification and aggregation of attitudes and opinions expressed by Internet users toward a specific topic. In this paper, we propose a pattern-based approach to detect sarcasm on Twitter. We propose four sets of features that cover the different types of sarcasm we defined. We use those to classify tweets as sarcastic and non-sarcastic. Our proposed approach reaches an accuracy of 83.1% with a precision equal to 91.1%. We also study the importance of each of the proposed sets of features and evaluate its added value to the classification. In particular, we emphasize the importance of pattern-based features for the detection of sarcastic statements.