A novel, information-theoretic fuzzy approach to discovering unreliable dat
a in a relational database is presented. A multilevel information-theoretic
connectionist network is constructed to evaluate activation functions of p
artially reliable database values. The degree of value reliability is defin
ed as a fuzzy measure of difference between the maximum attribute activatio
n and the actual value activation. Unreliable values can be removed from th
e database or corrected to the values predicted by the network. The method
is applied to a real-world relational database which is extended to a fuzzy
relational database by adding fuzzy attributes representing reliability de
grees of crisp attributes. The highest connection weights in the network ar
e translated into meaningful if, then rules. This work aims at improving re
liability of data in a relational database by developing a framework for di
scovering, accessing and correcting lowly reliable data. (C) 2001 Elsevier
Science B.V. All rights reserved.