Metric Categorization Relations Based on Support System Analysis

Abstract

Theoretical and practical aspects of classification systems and classification learning are considered. Analysis of subject area learning sets and analysis of classification schemes raises a number of nonstandard questions, such as relations between categorization/metadata and logic-combinatorial structuring/clustering of the descriptive part of the input table. Several such questions are are treated in this paper. Software environment of testing and evaluation is the extended system PaGaNe, completed by research methodology of well known logic-combinatorial scheme of pattern recognition.