CIC@PAN: Simplifying Irony Profiling using Twitter Data

The article explains the model submission by the team CIC for "Profiling Irony and Stereotype Spreaders on Twitter (IROSTEREO)" at PAN 2022. Irony profiling can help in identifying stereotype spreaders and can enhance the understanding of author behaviours. We proposed a methodology focusing on feature engineering to classify irony for long texts based on multiple linguistic and emotion-based features. We also extensively discussed the shortcomings of the data and the proposed task to provide the future research direction. The paper reveals the impact of robust feature engineering with a machine learning approach on the long social media texts in the author profiles. Our method achieved an accuracy of 87.22% on the test set.