Fuzzy Functional Dependency as a Sort of Semantic Knowledge: Representation, Preservation and Use

Abstract

Fuzzy functional dependency (FFD) reflects a sort of semantic knowledge about attributes. First, this paper presents general and specific forms of FFDs defined using fuzzy implication operators (FIOs), and describes the sufficient and necessary conditions for such FIOs to possess certain characteristics. Second, the paper discusses the preservation of FFDs associated with an original scheme as well as with the “simpler” schemes. Then a testing algorithm to see if any decomposition is dependency-preserving, and a computational algorithm to test whether a given set of FFDs is “partial-FFD-free” are provided. Third, the paper highlights two important areas FFDs can be used in fuzzy databases. One is that FFDs can serve as integrity constraints, which can be represented and enforced. The other is that FFDs play an important role in defining fuzzy normal forms, which help remedy possible data redundancy and update anomaly problems.Keywordsfuzzy databasesfuzzy functional dependency(FFD)dependency-preservationintegrity constraintsfuzzy normal forms