Fuzzy normal forms and a dependency-preserving decomposition into &thetas;-F3NF

In a fuzzy data model where imprecision of both attribute values and relationships among domain elements is taken into account, problems of data redundancy and update anomalies which are similar to those in classical databases are still found to exist. To cope with these problems, a number of fuzzy normal forms have been defined, based on fuzzy functional dependency, to serve as theoretical guidelines for fuzzy relational database design. This results in the notions of θ-keys, fuzzy first normal form (F1NF), θ-fuzzy second normal form (θ-F2NF), θ-fuzzy third normal form (θ-F3NF), and θ-fuzzy Boyce-Codd normal form (θ-FBCNF). Furthermore, the scheme decomposition into θ-F3NF is discussed in the light of the dependency-preserving property