Searching emotional scenes in TV programs based on twitter emotion analysis

Abstract

Twitter is a social networking service (SNS) that is specifically used to report the user's current status and what is going on in the presence of the user. One interesting new trend on Twitter is to tweet while watching a TV program. This paper proposes a method of analyzing emotions expressed in tweets. Our method assigns the emotional polarity values to tweets based on the dependency analysis as well as morphological analysis. The results of emotional analysis are used in indexing scenes in TV program viewer, and each scene is characterized with the emotions expressed in tweets posted at that time. This viewer allows users to search a TV program by referring to other Twitter users' emotional impressions for each scene. © 2013 Springer-Verlag Berlin Heidelberg.