Knowledge Predicates Recognition Systems Boolean Functions Logical Vectors
Issue Date:
2003
Publisher:
Institute of Information Theories and Applications FOI ITHEA
Abstract:
The concept of knowledge is the central one used when solving the various problems of data
mining and pattern recognition in finite spaces of Boolean or multi-valued attributes. A special form of
knowledge representation, called implicative regularities, is proposed for applying in two powerful tools of
modern logic: the inductive inference and the deductive inference. The first one is used for extracting the
knowledge from the data. The second is applied when the knowledge is used for calculation of the goal
attribute values. A set of efficient algorithms was developed for that, dealing with Boolean functions and finite
predicates represented by logical vectors and matrices.