Uncertainty in multi-source databases (Q1394747)

The book deals with the handling of supporting and opposing facts in multi-source databases. The author introduces the concept of confidence index sets (cisets), which is a generalization of fuzzy set theory. He uses tuples \((x,b,a)\), where \(x\) is a fact, \(b\) is a fuzzy set, representing the unfavourable factor, \(a\) stands for the favourable factor, both between 0 and 1. After an introduction with many illustrative examples the basic algebraic operations are presented. Second introduction leeds to an understanding of relational database systems with the definition of first normal form, keys and integrity constraints and a description of relational operators. In the third chapter the ciset relational model is presented and lateron extended to ciset relational algebra. As application to multi-source databases alternate worlds are used as tools to formalize the information content of a database with incomplete information. Usefullness of partial order on confidence indexes and additional operations are discussed in certain applications. In the last chapter the author summarizes the information source tracking as suggested by Sadri, which may leed to further extensions of ciset relational databases and applications. It ends with a probabilistic approach and the consideration of complexity. The author supplies the reader with many useful examples. All major results fundamental for this new approach to uncertainty in multi-sources relational databases are given.