Sentiment Detection for short english text (e.g. twitter messages), ranked on the 8th place out of the 50 participating submissions in the SemEval-2014 competition.
This is a joint project with Fatih Uzdilli and Mark Cieliebak from ZHAW.
Abstract: We describe a classifier to predict the message-level sentiment of English micro-blog messages from Twitter. This paper describes the classifier submitted to the SemEval-2014 competition (Task 9B). Our approach was to build up on the system of the last year’s winning approach by NRC Canada 2013 (Mohammad et al., 2013), with some modifications and additions of features, and additional sentiment lexicons. Furthermore, we used a sparse (l1-regularized) SVM, instead of the more commonly used l2-regularization, resulting in a very sparse linear classifier.
Update: Included our new paper on the 2015 competition as well. We again ranked in the top-ten of the 2015 competition (making us the only team placing in the top-ten both 2014 and 2015)!