Theory and Practice of Humor Computing for Social Media
Humor is a unique language expression of human beings,and people with a sense of humor can build a harmonious communication atmosphere.How to identify massive humor texts on social media and evaluate humor sense of users has become a major challenge in the field of computational linguistics.Therefore,from the perspective of cognitive linguistics,we propose a social media humor computing framework that covers resource construction,humor recognition,and humor sense evaluation.Firstly,in light of the semantic script theory of humor,we construct a Chinese humor corpus for social media.Then,based on semantic representation,combined with sentimental tendencies,semantic inconsistency,and fuzziness features,we recognize the humor on the text.Finally,based on obtaining the humor attributes and behavioral data of social media,we evaluate the user's sense of humor from the dimensions of creation and appreciation.Humor computing will help improve the level of language intelligence and enhance the application performance in areas such as reading comprehension,human-machine dialogue,and machine translation.
humor computingsocial media processingcorpus construction
NETL
NSTL
万方数据
