Mining approximate dependency to answer null queries on similarity-based fuzzy relational databases

Abstract

Null queries are queries that elicit a null answer from the database. In possibility-based fuzzy relational database model, a theoretical framework utilizing analogical reasoning and fuzzy functional dependency to answer null queries has been proposed by Dutta (1991). However, no searching algorithm is provided to discover the fuzzy functional dependencies among attributes. In this work, we extend the concept of fuzzy functional dependency to approximate dependency on similarity-based fuzzy relational data model. In addition, we proposed a data mining algorithm to discover all the approximate dependencies among attributes. It therefore can automatically obtain approximate answers for null queries and missing data values in an incomplete database. This kind of facility will certainly improve the cooperative nature of databases and enhance the user-friendliness of the database systems.