Experiments with DBpedia, WordNet and SentiWordNet as resources for sentiment analysis in micro-blogging

Second Joint Conference on Lexical and Computational Semantics (*SEM), Volume 2: Seventh International Workshop on Semantic Evaluation (SemEval 2013), pages 455–459, Atlanta, Georgia, June 14-15, 2013. c⃝2013 Association for Computational Linguistics Experiments with DBpedia, WordNet and SentiWordNet as re- sources for sentiment analysis in micro-blogging Abstract Sentiment Analysis in Twitter has become an important task due to the huge user-generated content published over such media. Such analysis could be useful for many domains such as Marketing, Finance, Politics, and So- cial. We propose to use many features in order to improve a trained classifier of Twitter mes- sages; these features extend the feature vector of uni-gram model by the concepts extracted from DBpedia, the verb groups and the similar adjectives extracted from WordNet, the Senti- features extracted using SentiWordNet and some useful domain specific features. We also built a dictionary for emotion icons, abbrevia- tion and slang words in tweets which is useful before extending the tweets with different fea- tures. Adding these features has improved the f-measure accuracy 2% with SVM and 4% with NaiveBayes. 1 Introduction In recent years, the explosion of social media has changed the relation between the users and the web. The world has become closer and more “real- time” than ever. People have increasingly been part of virtual society where they have created their content, shared it, interacted with others in differ- ent ways and at a very increasingly rate. Twitter is one of the most important social media, with 1 billion tweets1 posted per week and 637 million users2. 1http://blog.kissmetrics.com/twitter-statistics/ 2http://twopcharts.com/twitter500million.php With the availability of such content, it attracts the attention from who want to understand the opinion and interestingness of individuals. Thus, it would be useful in various domains such as poli- tics, financing, marketing and social. In this con- text, the efficacy of sentiment analysis of twitter has been demonstrated at improving prediction of box-office revenues of movies in advance of their release (Asur and Huberman, 2010). Sentiment Analysis has been used to study the impact of 13 twitter accounts of celebrated person on their fol- lowers (Bae and Lee, 2012) and for forecasting the interesting tweets which are more probably to be reposted by the followers many times (Naveed, Gottron et al., 2011). However, sentiment analysis of microblogs faces several challenges, the limited size of posts (e.g., maximum 140 characters in Twitter), the informal language of such content containing slang words and non-standard expressions (e.g. gr8 in- stead of great, LOL instead of laughing out loud, goooood etc.), and the high level of noise in the posts due to the absence of correctness verification by user or spelling checker tools. Three different approaches can be identified in the literature of Sentiment Analysis, the first ap- proach is the lexicon based which uses specific types of lexicons to derive the polarity of a text, this approach is suffering from the limited size of lexicon and requires human expertise to build the lexicon (Joshi, Balamurali et al., 2011). The second one is …

Content truncated to 500 words