AI-Based Learning Techniques for Sarcasm Detection of Social Media Tweets: State-of-the-Art Survey
- PDF / 1,150,165 Bytes
- 14 Pages / 595.276 x 790.866 pts Page_size
- 19 Downloads / 189 Views
SURVEY ARTICLE
AI‑Based Learning Techniques for Sarcasm Detection of Social Media Tweets: State‑of‑the‑Art Survey Yogesh Kumar1,2 · Nikita Goel1,2 Received: 4 August 2020 / Accepted: 17 September 2020 © Springer Nature Singapore Pte Ltd 2020
Abstract Sarcasm, though difficult to define but plays a crucial role in one’s life. Sarcasm as a jest is a matter of fun but when taken seriously can cause unwelcoming results. Sometimes, sarcasm is defined as “a sharp, bitter, or cutting expression or remark; a bitter jibe or taunt”. These days’ researchers are working towards the detection of sarcasm for the purpose of sentiment analysis. Emotion and sentiment-bearing information are carried by subjective sarcastic sentences. The objective of the paper is to highlight the different types of sarcastic tweets and their usage in sentiment analysis. The authors mainly emphasize several approaches which include sentiment analysis, machine and deep learning classifications. The paper focuses on the use of machine learning and deep learning for identifying sarcastic tweets. Numerous feature extraction techniques have been studied and machine and deep learning classifications have been taken into account. The comparative table shows the results obtained using the various evaluation metrics such as accuracy, precision, recall, and f-score. Keywords Feature vector · N-gram · Sentiment analysis · Sarcasm detection · Machine learning · Deep learning
Introduction Today’s world is data driven, and the credit goes to technological advancements in the field of communication like mobile phones, social websites and many more. Such advancements have led to exponential increase in data generation. Recent years, we have witnessed tremendous usage in social websites such as Twitter, Facebook where people come together and share their thoughts, opinions, discoveries and also engage in various discussions. It is necessary to analyze this data for various purposes such as sentiment analysis, judging the tone of the writer, and many more. It is necessary to understand the mood of the writer who adds This article is part of the topical collection “Computational Statistics” guest edited by Anish Gupta, Mike Hinchey, Vincenzo Puri, Zeev Zalevsky and Wan Abdul Rahim. * Yogesh Kumar [email protected] Nikita Goel [email protected] 1
Chandigarh Group of Colleges, Landran, Mohali, India
Maharaja Agrasen Institute of Technology, GGSIPU, Delhi, India
2
data to such social websites as these data have the capability to influence the mob. The mood is polymorphic in nature, ranging from being puzzled to provoke or diverted to nauseate. It is a subject of research among psychologists to study different moods of people and their origin. Moods have an influential effect on one’s behavior which can affect not only their lives but of others too. Moods are related to the feelings and concentrated mainly on opinion and attitude. That is why sentiments are considered to be subjective in nature. Some people refer to feelings as a natural w
Data Loading...
