• DocumentCode
    1167767
  • Title

    An efficient indefiniteness inference scheme in indefinite deductive databases

  • Author

    Ku, Cyril S. ; Kim, Heung D. ; Henschen, Lawrence J.

  • Author_Institution
    Bellcore, Piscataway, NJ, USA
  • Volume
    6
  • Issue
    5
  • fYear
    1994
  • fDate
    10/1/1994 12:00:00 AM
  • Firstpage
    713
  • Lastpage
    722
  • Abstract
    We introduce an inference scheme, based on the compilation approach, that can answer “true,” provable-false,” “indefinite,” or “assumable-false” to a closed query in an indefinite deductive database under the generalized closed world assumption. The inference scheme proposed in this paper consists of a representation scheme and an evaluation process that uses one of two groups of positive indefinite ground clauses (PIGC´s) derivable from the database for a given query. These two groups of PIGC´s are base-PIGC´s and descendants of base-PIGC´s. We prove that the set of base-PIGC´s derivable from the database is sufficient to infer the indefiniteness of a query. This newly proposed method offers many advantages over the existing compilation method. This new method outperforms the existing one in terms of the cost of compilation, ease in the handling of updates, and efficiency in query evaluation
  • Keywords
    database theory; deductive databases; inference mechanisms; knowledge representation; query processing; uncertainty handling; closed query; compilation approach; cost; generalized closed world assumption; indefinite deductive database; indefinite deductive databases; indefiniteness inference scheme; knowledge representation; positive indefinite ground clauses; query evaluation; update handling; Algebra; Computer science; Costs; Database systems; Deductive databases; Information retrieval; Query processing; Relational databases; Testing;
  • fLanguage
    English
  • Journal_Title
    Knowledge and Data Engineering, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    1041-4347
  • Type

    jour

  • DOI
    10.1109/69.317702
  • Filename
    317702