IIT Delhi at SemEval-2018 Task 1 : Emotion Intensity Prediction

Proceedings of the 12th International Workshop on Semantic Evaluation (SemEval-2018), pages 339–344 New Orleans, Louisiana, June 5–6, 2018. ©2018 Association for Computational Linguistics IIT Delhi at SemEval-2018 Task 1 : Emotion Intensity Prediction Bhaskar Kotakonda Undegraduate IIT Delhi ee1140448@iitd.ac.in Prashanth Gowda Undergraduate IIT Delhi ee1140470@iitd.ac.in Brejesh Lall Associate Professor IIT Delhi brejesh@ee.iitd.ac.in Abstract This paper discusses the experiments per- formed for predicting the emotion intensity in tweets using a generalized supervised learning approach. We extract 3 kind of features from each of the tweets - one denoting the senti- ment and emotion metrics obtained from dif- ferent sentiment lexicons, one denoting the se- mantic representation of the word using dense representations like Glove, Word2vec and fi- nally the syntactic information through POS N-grams, Word clusters, etc. We provide a comparative analysis of the significance of each of these features individually and in com- bination tested over standard regressors avail- able in scikit-learn. We apply an ensemble of these models to choose the best combination over cross validation. Our resources and the details of implementa- tion are publicly available at : https://github.com/prashanth470/ affect-in-tweets 1 Introduction In Natural Language Understanding, the field of sentiment analysis deals with the process of deter- mining the polarity of a given text, such as pos- itive, negative, neutral and mixed. In extension to this analysis, we have the emotion recognition task which deals with associating the text with pre- defined set of emotions like anger, fear, joy, etc. A general method of performing the emotion recog- nition task is to employ weak supervision mod- els like emojis, hashtags and emoticons to mine emotion. Instead of using this discreet approach to emotion, continuous models that map text to an n - dimensional space with valence, arousal and dominance can be used. Another interesting problem in the NLP space is the abundance of social media texts, especially twitter. Twitter is a micro-blogging site where people express themselves and react to content in real-time. An estimated 500 million tweets are generated on a daily basis. The peculiar nature of such micro-blogging sites is the form of expres- sion through hashtags, emojis, slang and informal words etc. But analyzing this abundant informa- tion would help us to realize several insights about an event, person or organization. It is with this motivation that the SemEval shared task on Emotion Intensity was con- ducted.(Mohammad et al., 2018) Given a tweet and an emotion (anger, fear, sadness, joy) the aim is to determine the intensity score that can be seen as an approximation to the intensity felt by the reader or expressed by the author. The paper is divided into 3 sections hereon - the second sec- tion talks about the system description, the third section on a comparative analysis of results and fi- nally a discussion on the future scope. 2 System Description The datasets for anger, joy, fear and sad- ness were created using a technique called the Best Worst Scaling.(Mohammad and Kiritchenko, 2016) These annotations lead to reliable fine grained …

Content truncated to 500 words