Social Media and Personal Data
     • Much personal information 
        revealed in social media
         –Content, links, ratings  
           personal preferences
     • All this information is useful 
        to
         –Researchers: social science
         –Businesses: targeted 
           advertising
      Dec 5, 2014                     AIRS 2014                             2
           User Biographies in Twitter
     • Self-introductions written in free form
     • Reflect users’ background and interests
      Dec 5, 2014                     AIRS 2014                             3
              User Biographies in Twitter
                  age             profession                interests
      Around 28% of Singapore Twitter users and 50% of US 
      Twitter users
      revealed their personal interests in their biographies.
                               Dong Wei et. al. Who am I on Twitter?: A cross-country comparison.
                               WWW’2014
       Dec 5, 2014                             AIRS 2014                                      4
                                 Outline
     • Background
     • Our task
     • Syntactic patterns of interest tags
     • Build training data + gold standard
     • Method
     • Experiments
     • Summary
      Dec 5, 2014                     AIRS 2014                             5
                                Our task
     • Automatically extract phrases that 
        describe a user’s personal interests.
         –We call them “interest tags”
         –A typical information extraction problem.
         –Automatically build training data based on 
            common syntactic patterns.
      Dec 5, 2014                     AIRS 2014                             6