TF-IDF Feature Extraction based Sarcasm Detection on Social Media

Abstract

The popularity of social media has significantly changed how people live their everyday. Twitter is being the most widely used platform for information sharing and emotional expression. However, the excess of false and controversial information has raised concerns as social media usage increases. One problem in particular is the prevalence of sarcastic text, which can be challenging to identify manually. In proposed research, an innovative approach is developed that makes use of cutting-edge machine learning methodologies for sarcasm detection on social media. The proposed approach includes four machine learning algorithms for classification and the TF-IDF (Term Frequency- Inverse Document Frequency) to extract features from the text. Different evaluation metrics like recall, precision, F1, and accuracy scores were used to evaluate the highest values of the various algorithms.