SINAI: Machine Learning and Emotion of the Crowd for Sentiment Analysis in Microblogs

Second Joint Conference on Lexical and Computational Semantics (*SEM), Volume 2: Seventh International Workshop on Semantic Evaluation (SemEval 2013), pages 402–407, Atlanta, Georgia, June 14-15, 2013. c⃝2013 Association for Computational Linguistics SINAI: Machine Learning and Emotion of the Crowd for Sentiment Analysis in Microblogs E. Mart´ınez-C´amara SINAI research group University of Ja´en E-23071, Ja´en (Spain) emcamara@ujaen.es A. Montejo-R´aez SINAI research group University of Ja´en E-23071, Ja´en (Spain) amontejo@ujaen.es M. T. Mart´ın-Valdivia SINAI research group University of Ja´en E-23071, Ja´en (Spain) maite@ujaen.es L. A. Ure˜na-L´opez SINAI research group University of Ja´en E-23071, Ja´en (Spain) laurena@ujaen.es Abstract This paper describes the participation of the SINAI research group in the 2013 edi- tion of the International Workshop Se- mEval. The SINAI research group has submitted two systems, which cover the two main approaches in the field of sen- timent analysis: supervised and unsuper- vised. 1 Introduction In the last years, the sentiment analysis (SA) re- search community wants to go one step further, which consists in studying different texts that usually can be found in commerce websites or opinions websites. Currently, the users publish their opinions through other platforms, being one of the most important the microblogging plat- form Twitter1. Thus, the SA research commu- nity is focused on the study of opinions that users publish through Twitter. This interest is shown in several workshops focused on the study of SA in Twitter: 1. RepLab 2012 at CLEF2 (Amig´o et al., 2012): Competition carried out within the CLEF conference, where the participants had to develop a system for measuring the reputation of commercial brands. 1http://twitter.com 2http://limosine-project.eu/events/ replab2012 2. TASS 2012 at SEPLN3(Villena-Rom´an et al., 2013): Satellite event of the SEPLN 2012 Conference to foster the research in the field of SA in social media, specifically focused on the Spanish language. In this paper is described the participation of the SINAI4 research group in the second task of the 2013 edition of the International Workshop SemEval (Wilson et al., 2013). We have submit- ted two systems (constrained and unconstrained). The constrained system follows a supervised ap- proach, while the unconstrained system is based on an unsupervised approach which used two lin- guistic resources: the Sentiment Analysis Lexi- con5 (Hu and Liu, 2004) and WeFeelFine6 (Kam- var and Harris, 2011). The paper is organized as follows: first we present a description of the preparing data pro- cess. Then the constrained system is outlined. The participation overview finishes with the de- scription of the unconstrained system. 2 Preparing data The organizers provided two sets of data, one for training and another for the development. The data was concerned by a set of identification number of tweets with their corresponding po- larity label. We used the script provided by the organizers to download the two sets of tweets. 3http://www.daedalus.es/TASS/ 4http://sinai.ujaen.es 5http://www.cs.uic.edu/˜liub/FBS/ opinion-lexicon-English.rar 6http://wefeelfine.org 402 The python script was no able to download all the tweets. The training set was composed by 8,633 tweets and the development set by 1,053 tweets. The data preparation is a step in …

Content truncated to 500 words