医療におけるアクティブマイニング : Medical Data Miningからの新たな展開(特別講演,医療情報システムにおけるソフトコンピューティグ)

抄録

It has been pointed out that whether knowledge is discovered by data mining methods strongly depends on the deep interpretation of the patterns obtained in which domain knowledge plays important roles. Thus, it is highly expected that incorporation of domain knowledge into KDD process is indispensable to supporting the interpretation process. In this paper, first, the importance of pre- and postprocessing is illustrated in knowledge discvoery from examples in medical domain. Then, the extension of KDD : active mining is proposed for actual discovery in medical domain.