Sarcasm detection of non # tagged statements using MLP-BP
Pornima Gidhe, Leena R. Ragha · 2017
These days being sarcastic on social site is common for everyone. Sarcasm is used to say something which is opposite of original literal meaning. Sarcasm is used by people to show anger, joy, contradict, humour etc but people frame the sarcasm to hide actual meaning i.e. semantics and emotions. Sarcasm is defined as encapsulating sarcastic features in literal meaning thus we can say extracting sarcastic features is a key part of sarcasm detection. Extracting sarcastic features from such text is a challenge on which researchers are working. Researchers worked on dataset having # sarcasm tag using structural and sentiment features. There are some cases where # sarcasm tags are not available, to disambiguate such cases emotional and semantic similarity features are required. We propose technique to classify sarcastic and non-sarcastic sentences with the help of structural, affective and semantic similarity features using MLP-BP.