Learning Sentiments from Tweets with Personal Health Information

Abstract

We present results of sentiment analysis in Twitter messages that disclose personal health information. In these messages (tweets), users discuss ailment, treatment, medications, etc. We use the author-centric annotation model to label tweets as positive sentiments, negative sentiments or neutral. The results of the agreement among three raters are reported and discussed. We then use Machine Learning methods on multi-class and binary classification of sentiments. The obtained results are comparable with previous results in the subjectivity analysis of user-written Web content.