Emoji-Enhanced Sarcasm Detection: A Comparitive Study of Logistic Regression and LSTM Models

Emoji-Enhanced Sarcasm Detection: A Comparitive Study of Logistic Regression and LSTM Models

Authors- Aamina Atlaswala, Dr. Rakhi Gupta, Nashrah Gowalker

Abstract-Sarcasm is a form of language defined using words or phrases that express the opposite of their actual meaning, often with a purpose of mocking or criticizing something or someone. Sarcasm detection is crucial for mining, sentiment analysis, detecting cyberbullies, online trolls, and other similar activities. Detecting Sarcasm is a part of Sentimental Analysis. This research focuses on detecting sarcasm in text with emojis using the Logistic Regression Model, a traditional machine learning approach, and the Long Short- Term Memory (LSTM) Model a deep learning technique and provides a comparative study of models from the two branches of Artificial Intelligence. The emojis in the dataset were converted to their text descriptions to infer their actual usage in the text. And then the models were trained on the preprocessed data.

DOI: /10.61463/ijset.vol.12.issue5.300