Abstract: Humor is the most gripping factor in human to human communication which provokes laughter and or amusement. It also have a great impact on human beings mental and physical health and hence, it is an essential element in communication. The purpose of this paper is to survey different approaches for humor recognition. Various methodologies are proposed for recognition of humor and they are Machine Learning, Discourse Parsing and Deep Learning. The paper also discuss the advantages and limitations of methodologies used. The various datasets used for Humor recognition is also discussed in this paper.

Keywords: Convolutional Neural Network (CNN),  Humor Recognition,  Long Short-Term Memory (LSTM), Rhetorical Structure Theory (RST)