Sentence Patterns based Sarcasm Detection and Classification using Long Short Term Memory Model

Nowadays, an individual's opinion or views are expressed on social media. These opinions could be sarcastic or ironic, which leads to the rise of various sentiments. Most organizations aim to leverage these views in order to improve their business. This gave rise to sentimental analysis. This paper focuses on how sarcasm can be classified using sentence structures. Sentences are extracted from social media tweets. The sentences are divided into three categories: simple, complex, and mixed. A Long-Short Term Memory (LSTM), along with GLoVe embeddings, is used to classify the sentences as sarcastic or non-sarcastic