PPT Slide

Some KDD work on clustering (KDD’95 ~ KDD’98)

• Partition-based
  　Scaling k-means, EM clustering for large
  
  databases (Bradley, Fayyad, Reina; Microsoft)

• Model-based
  　Baysian clustering AUTOCLASS. Discovered new class of
  
  galaxies infra-red satellite data in which astronomers
  could not see the classes (Cheeseman et al., NASA).

• Mathematical programming

　　　in　formulating k-means, k-medians. Applied to the Wisconsin

　Breast Cancer Dataset (Mangasarian, Wisconsin Univ.)

前のｽﾗｲﾄﾞ

次のｽﾗｲﾄﾞ

最初のｽﾗｲﾄﾞに戻る

ｸﾞﾗﾌｨｯｸｽの表示