‘Sentiment mining’ – i.e. trying to gauge the Public’s attitude towards an institution, product, firm (etc. etc.) though automatic analysis of Social Media posts (etc. etc.) is now considered an essential tool for market researchers and ‘reputation managers’.
But there are problems. One of which is sarcasm. Given its prevalence, serious errors can be introduced in a Sentiment Mining Picture if it’s not reliably detected.
Work on Automatic Sarcasm Detection began in the mid 2000s – see, for example, Joseph Tepperman, David Traum, and Shrikanth S. Narayanan. “Yeah right”: Sarcasm recognition for spoken dialogue systems. In Proceedings of InterSpeech, pp. 1838–1841, Pittsburgh, PA, sep 2006.
Since then, work on refining Automatic Sarcasm Detection Algorithms has flourished. Here is a list of but-a-few of the many hundreds of scholarly works which address the issue (in no particular order).
● An Approach to Detect Sarcasm in Tweets
● Sarcasm detection in mash-up language using soft-attention based bi-directional LSTM and feature-rich CNN
● Sarcasm Detection Using Feature-Variant Learning Models
● A Review Paper on Sarcasm Detection
● Sarcasm Detection with Sentiment Semantics Enhanced Multi-level Memory Network
● Sarcasm Detection Methods in Deep Learning: Literature Review
● Sarcasm Detection Using Multi-Head Attention Based Bidirectional LSTM
● Sarcasm Detection Using Deep Learning-Based Techniques
● A Multi-Dimension Question Answering Network for Sarcasm Detection
● Tweet sarcasm detection using deep neural network
● Sarcasm detection using Grice’s maxims
● Modelling Context with User Embeddings for Sarcasm Detection in Social Media’
● ‘The perfect solution for detecting sarcasm in tweets #not’
Note: One of the simplest methods, pointed out in An Approach to Detect Sarcasm in Tweets (amongst other papers above) is to look for posts or Tweets (etc. etc.) which use the hashtag #Sarcasm.
Research research by Martin Gardiner