Applications of Nonclassical Logic Methods for Purposes of Knowledge Discovery and Data Mining

Abstract

Methods for solution of a large class of problems on the base of nonclassical, multiple-valued, and probabilistic logics have been discussed. A theory of knowledge about changing knowledge, of defeasible inference, and network approach to an analogous derivation have been suggested. A method for regularity search, logic-axiomatic and logic-probabilistic methods for learning of terms and pattern recognition in the case of multiple-valued logic have been described and generalized. Defeasible analogical inference and new forms of inference using exclusions are considered. The methods are applicable in a broad range of intelligent systems.