Mathematical Analysis of Different Learning Approaches on User Behavior and Contextual Evaluation for Sarcasm Prediction

Abstract

A large number of people have been using social media platforms extensively to communicate their thoughts and feelings in the recent era of social networking. Both the user base and data volume on social networks are growing quickly every day. Any time an event or activity occurs nearby, nearby individuals express their thoughts and reactions on social media. When a new product is introduced, users on social media platforms also comment on it. Some people express their views or feelings using informal or complex language which makes it difficult to understand for another user. It is challenging to ascertain the true thoughts because different people express their opinions in complex ways. In this study, the various factors that affect these feelings are briefly discussed. In order to identify sarcasm on Twitter, a generic technique is also necessary in addition to the tweet's content. The proposed approach uses contents of tweet in association with important aspects like user behavior and context of tweet. By users’ behavior we can identify its influence on other users and context is required to identify user behavior while detecting sarcasm. Proposed approach uses user behavior pattern and personality features along with contextual data. This all information and the already known sarcasm prediction mechanism will help us to set up the generic approach to detect sarcasm on Twitter.