Systematic Literature Review: Computational Approaches for Humour Style Classification

Mary Ogbuka Kenneth, Foaad Khosmood, Abbas Edalat

Abstract:

Understanding humour styles is crucial for grasping humour’s diverse nature and impact on psychology and artificial intelligence. Humour can have both therapeutic and harmful effects, depending on the style. Though computational humour style analysis studies are limited, extensive research exists, particularly in binary humour and sarcasm recognition. This systematic literature review explores computational techniques in these related tasks, revealing their relevance to humour style analysis. It uncovers common approaches, datasets, and metrics, addressing research gaps. The review identifies features and models that can transition smoothly from binary humour and sarcasm recognition to humour style identification. These include incongruity, sentiment analysis, and various models, like neural networks and transformer-based
models. Additionally, it provides access to humour-related datasets, aiding future research.