Sarcasm Detection in Text Data Using Glove Embedding

Abstract

Sarcasm is a way of expressing feelings in which people say or write something which is completely different or opposite to what they actually mean to say. Hence it is very difficult to identify sarcasm . It is usually an ironic or satirical remark tempered by humor. Mainly, people use it to say the opposite of what's true to make someone look or feel foolish. Understanding the sarcasm can improve the accuracy of sentiment analysis. Sentiment analysis (or opinion mining) is a natural language processing technique used to determine whether data is positive, negative or neutral. This helps in identifying what the opinions of users or individual or society are. In this project an attempt is made to develop a model to detect if a sentence is sarcastic or if it is not sarcastic. Keywords: Sarcasm detection, GloVe Embedding, LSTM, Natural Language Processing, Sentiment