Dataset 4 sub-datasets derived from 20-newsgroups and rcv1 Document datasets 20NewsGroup Sub-datasets #(Document) word) #(entity) #(types) 20NG-SIM 3,000 8,010 11,192 219 20NG-DIF 3,000 9,182 13,297 251 GCAG-SIM 3,596 11,096 0.540 022 GCAT-DIF 2,700 13,291 13,179 261 RCV1-GCAT Each sub-datasets consists of three similar or distinct topics 27Dataset • 4 sub-datasets derived from 20-newsgroups and RCV1 Document datasets Sub-datasets #(Document) #(word) #(Entity) #(Types) 20NG-SIM 3,000 8,010 11,192 219 20NG-DIF 3,000 9,182 13,297 251 GCAG-SIM 3,596 11,096 10,540 233 GCAT-DIF 2,700 13,291 13,179 261 Each sub-datasets consists of three similar or distinct topics. 20NewsGroup RCV1-GCAT 27