IITPatna: Supervised Approach for Sentiment Analysis in Twitter

Proceedings of the 8th International Workshop on Semantic Evaluation (SemEval 2014), pages 324–328, Dublin, Ireland, August 23-24, 2014. IIT Patna: Supervised Approach for Sentiment Analysis in Twitter Raja Selvarajan and Asif Ekbal Department of Computer Science and Engineering Indian Institute of Technology Patna, India {raja.cs10,asif}@iitp.ac.in Abstract In this paper we report our works for SemEval-2014 Sentiment Analysis Twitter evaluation challenge. This is thein first time we attempt for this task, and our submissions are based on supervised machine learning algorithm. We use Sup- port Vector Machine for both the tasks, viz. contextual polarity disambiguation and message polarity classification. We identify and implement a small set of features for each the tasks, and did not make use of any external resources and/or tools. The systems are tuned on the devel- opment sets and finally blind evaluation is performed on the respective test set, which consists of the datasets of five different domains. Our submission for the first task shows the F-score values of 76.3%, 77.04%, 70.91%, 72.25% and 66.32% for LiveJournal2014, SMS2013, Twitter2013, Twitter2014 and Twitter2014Sarcasm datasets, respectively. The system devel- oped for the second task yields the F-score values of 54.68%, 40.56%, 50.32%, 48.22% and 36.73%, respectively for the five different test datasets. 1 Introduction During the past few years, the communications in the forms of microblogging and text messaging have emerged and become ubiquitous. Opinions and sentiments about the surrounding worlds are widely expressed through the mediums of Twit- ter messages (Tweets) and Cell phone messages (SMS). The availability of social content gener- ated on sites such as Twitter creates new opportu- This work is licensed under a Creative Commons At- tribution 4.0 International Licence. Page numbers and pro- ceedings footer are added by the organisers. Licence details: http://creativecommons.org/licenses/by/4.0/ nities to automatically study public opinion. Deal- ing with these informal text genres presents new challenges for data mining and language process- ing techniques beyond those encountered when working with more traditional text genres such as newswire. Tweets and SMS messages are short in length, usually a sentence or a headline rather than a document. These texts are very informal in nature and contains creative spellings and punctu- ation symbols (Nakov et al., 2013). Text also con- tains lots of misspellings, slang, out-of-vocabulary words, URLs, and genre-specific terminology and abbreviations, e.g., RT for reTweet and #hash- tags. The kind of these specific features pose great challenges for building various lexical and syntac- tic resources and/or tools, which are required for efficient processing of texts. These aspects also introduce complexities to build the state-of-the- art data mining systems. In recent times, there has been a huge interest to mine and understand the opinions and sentiments that people are com- municating in social media (Barbosa and Feng, 2010; Bifet et al., 2011; Pak and Paroubek, 2010; Kouloumpis et al., 2011). Recent studies show the interests in sentiment analysis of Tweets across a variety of domains such as commerce (Jansen et al., 2009), health (Chew and Eysenbach, 2010; Salathe and Khandelwal, 2011) and …

Content truncated to 500 words