Sarcasm Detection in English Text using Tweets and Headlines

Online sarcasm is usually used to express something opposite to the literal meaning of the phrase or sentence and is hard to catch since it thrives in ambiguous situations. This paper revolves around natural language processing of samples in the English language to generalize the structure of a sarcastic sentence. Through this study, we develop a well-defined model for the automatic identification of sarcastic sentences in various usages. We intend for this study to find use in opinion mining, information categorization and sentiment analysis. Through the LSTM model along with a plethora of data used, this study achieved an accuracy of 93.02% without overfitting.