Enhanced sentiment learning using Twitter hashtags and smileys

Automated identification of diverse sen-timent types can be beneficial for many NLP systems such as review summariza-tion and public media analysis. In some of these systems there is an option of assign-ing a sentiment value to a single sentence or a very short text. In this paper we propose a supervised sentiment classification framework which is based on data from Twitter, a popu-lar microblogging service. By utilizing 50 Twitter tags and 15 smileys as sen-timent labels, this framework avoids the need for labor intensive manual annota-tion, allowing identification and classifi-cation of diverse sentiment types of short texts. We evaluate the contribution of dif-ferent feature types for sentiment classifi-cation and show that our framework suc-cessfully identifies sentiment types of un-tagged sentences. The quality of the senti-ment identification was also confirmed by human judges. We also explore dependen-cies and overlap between different sen-timent types represented by smileys and Twitter hashtags.