Research Paper
An efficient approach for sarcasm detection in tweets using polarity flip
Sentiment analysis is the qualitative method that uses natural language processing, text analysis, and computational linguistics. The goal of sentiment analysis is to determine if a specific passage in the text shows positive, negative or neutral sentiment towards the subject. Social media platforms, like Twitter, offer a wide ability that allows users to express their thoughts by employing figurative language devices such as sarcasm to achieve different communication purposes. Dealing with such kind of content represents a big challenge for computational linguistics. Due to these difficulties and the inherently tricky nature of sarcasm, it is generally ignored during social network analysis. As a result, the outcome of such analysis is affected adversely. Also, nowadays, Customers use twitter as a platform to express their emotions and opinions about social issues, products or services. Lexicon based methods of text mining can at times fail to recognize sarcasm used by customers online. This has direct implications for companies using text mining on online content to identify new customers, address customer problems, and reduce customer turn rates or for any other CRM (Customer Relationship Management) activities. Thus, sarcasm detection poses to be one of the most critical problems which we need to overcome while trying to yield high accuracy insights from abundantly available data. This work considers sarcastic tweets (including those based on products) and proposes an effective approach to detect sarcasm. The system utilizes 21 features based on context, contrast, and emotions. The results show that sarcastic tweets inclined to have more polarity flips than Non-sarcastic tweets. Also, it is found that MLP and the Random Forest classifier tend to perform better than other classifiers with an accuracy of 94%.
Published by: Karthika K.
Author: Karthika K.
Paper ID: V5I2-2114
Paper Status: published
Published: April 26, 2019
Full Details