These data files contain the id as well as the target of the tweets as described in the paper "Sarcastic or Not: Word-Embeddings to Predict the Literal or Sarcastic Meaning of Words". The first column represents the target and the second column represents the id of the tweet.
The file names show the purpose of the file; for instance, "tweet.SARCASM.all.id.TRAIN" contains sarcastic training data for all the targets used in the paper (37 targets) where as "tweet.SENTIMENT.all.id.TEST" and "tweet.NON_SARCASM.all.id.TEST" contain the sentiment test data and random test data, respectively.
Please cite the following paper if you write a research paper using this data.
Sarcastic or Not: Word-Embeddings to Predict the Literal or Sarcastic Meaning of Words. Debanjan Ghosh, Weiwei Guo, Smaranda Muresan. In Proceedings of EMNLP, 2015, Lisbon, Portugal.
Please contact Debanjan Ghosh (debanjan.ghosh@rutgers.edu) if you encounter any problems.