Finding Social Relationships by Extracting Polite Language in Micro-blog Exchanges

The aim of this study was to describe user relationships based on honorific expressions in messages posted to a micro-blogging service and to classify the users into appropriate groups. In particular, we focused on attitudinal expressions that indicate the speaker’s attitude. We compiled posts on micro-blogging platform Twitter and performed an experiment to classify the data based on honorifics. In the results, compared with indegree centrality values, the obtained social graph was superior to one acquired from a baseline, i.e., by the condition of the follower-followed relationship.