SVNIT @ SemEval 2017 Task-6: Learning a Sense of Humor Using Supervised Approach

Proceedings of the 11th International Workshop on Semantic Evaluations (SemEval-2017), pages 411–415, Vancouver, Canada, August 3 - 4, 2017. c⃝2017 Association for Computational Linguistics SVNIT @ SemEval 2017 Task-6: Learning a Sense of Humor Using Supervised Approach Rutal Mahajan, Mukesh Zaveri Computer Engineering Department S.V. National Institute of Technology, Surat {rutal.mahajan, mazaveri}@gmail.com Abstract This paper describes the system developed for SemEval 2017 task 6: #HashTagWars - Learning a Sense of Humor. Learning to recognize sense of humor is the important task for language understanding applica- tions. Different set of features based on frequency of words, structure of tweets and semantics are used in this system to identify the presence of humor in tweets. Supervised machine learning approaches, Multilayer perceptron and Naïve Bayes are used to classify the tweets in to three lev- els of sense of humor. For given Hashtag, the system finds the funniest tweet and predicts the amount of funniness of all the other tweets. In official submitted runs, we have achieved 0.506 accuracy using multi- layer perceptron in subtask-A and 0.938 distance in subtask-B. Using Naïve bayes in subtask-B, the system achieved 0.949 distance. Apart from official runs, this sys- tem have scored 0.751 accuracy in sub- task-A using SVM. 1 Introduction Humor is an integral aspect of human beings that requires self-awareness, spontaneity, linguistic sophistication and empathy. Generating and rec- ognizing humor is not an easy task to be carried out by machines. Generating and understanding humor can be useful in many NLP tasks. (Azizinezhad & Hashemi, 2011) have described the use of humor as the pedagogical tool for lan- guage learners, as it helps to keep students inter- ested and motivated. Moreover, recognizing hu- mor is also important in sentiment analysis and opinion mining because it can be useful to get the actual meaning out of figurative sentence. Research on modeling humor such as (Barbieri & Saggion, 2014)(Raz, 2012)is focused on classi- fying humor into binary classes as humor and non-humor. In (Reyes, Rosso, & Buscaldi, 2012), humor is modeled by a binary classifier as well as by a multi-class classifier. It classifies different figurative sentences into humor, irony, politics, technology and general sentences. But all these approaches ignore the continuous nature of hu- mor. Hence in task 6 of SemEval 2017 HashTag Wars: Learning a sense of humor (Potash, Romanov, & Rumshisky,2016), humor in tweet should be modeled in its continuous form instead of binary. The participating groups are asked to predict the amount of funniness of the tweet for particular hashtag according to gold labels of tweet. Tweets are labeled with 0, 1, or 2. 0 corre- sponds to tweet not in top 10. 1 corresponds to tweet in top 10 but not winning tweet and 2 corre- sponds to winning tweet. There are two subtasks: A) pairwise comparison- a task of predicting which tweet is funnier from given two tweets ac- cording to gold labels of tweets. In given pair of tweets, the tweet with higher label is said to be funnier. …

Content truncated to 500 words