Biocom Usp: Tweet Sentiment Analysis with Adaptive Boosting Ensemble

Proceedings of the 8th International Workshop on Semantic Evaluation (SemEval 2014), pages 129–134, Dublin, Ireland, August 23-24, 2014. Biocom Usp: Tweet Sentiment Analysis with Adaptive Boosting Ensemble N´adia F. F. Silva, Eduardo R. Hruschka University of S˜ao Paulo, USP S˜ao Carlos, SP, Brazil nadia, erh@icmc.usp.br Estevam Rafael Hruschka Jr. Department of Computer Science Federal University of Sao Carlos. S˜ao Carlos, SP, Brazil estevam@dc.ufscar.br Abstract We describe our approach for the SemEval-2014 task 9: Sentiment Analy- sis in Twitter. We make use of an en- semble learning method for sentiment classification of tweets that relies on varied features such as feature hash- ing, part-of-speech, and lexical fea- tures. Our system was evaluated in the Twitter message-level task. 1 Introduction The sentiment analysis is a field of study that investigates feelings present in texts. This field of study has become important, espe- cially due to the internet growth, the content generated by its users, and the emergence of the social networks. In the social networks such as Twitter people post their opinions in a colloquial and compact language, and it is be- coming a large dataset, which can be used as a source of information for various automatic tools of sentiment inference. There is an enor- mous interest in sentiment analysis of Twit- ter messages, known as tweets, with applica- tions in several segments, such as (i) directing marketing campaigns, extracting consumer re- views of services and products (Jansen et al., 2009); (ii) identifying manifestations of bully- ing (Xu et al., 2012); (iii) predicting to fore- cast box-office revenues for movies (Asur and Huberman, 2010); and (iv) predicting accep- tance or rejection of presidential candidates (Diakopoulos and Shamma, 2010; O’Connor et al., 2010). This work is licensed under a Creative Commons Attribution 4.0 International Li- cence. Page numbers and proceedings footer are added by the organisers. Licence details: http://creativecommons.org/licenses/by/4.0/ One of the problems encountered by re- searchers in tweet sentiment analysis is the scarcity of public datasets. Although Twit- ter sentiment datasets have already been cre- ated, they are either small — such as Obama- McCain Debate corpus (Shamma et al., 2009) and Health Care Reform corpus (Speriosu et al., 2011) or big and proprietary such as in (Lin and Kolcz, 2012). Others rely on noisy labels obtained from emoticons and hashtags (Go et al., 2009). The SemEval-2014 task 9: Sen- timent Analysis in Twitter (Nakov et al., 2013) provides a public dataset to be used to com- pare the accuracy of different approaches. In this paper, we propose to analyse tweet sentiment with the use of Adaptive Boost- ing (Freund and Schapire, 1997), making use of the well-known Multinomial Classi- fier. Boosting is an approach to machine learning that is based on the idea of creat- ing a highly accurate prediction rule by com- bining many relatively weak and inaccurate rules. The AdaBoost algorithm (Freund and Schapire, 1997) was the first practical boost- ing algorithm, and remains one of the most widely used and studied, with applications in numerous fields. Therefore, …

Content truncated to 500 words