Decision analysis

Reducing dimensionality of attribute space: set encapsulation

A. B. Petrovskii

Federal Research Center "Computer Science and Control" of Russian Academy of Sciences, Moscow

Abstract: The paper describes a new mathematical apparatus for set encapsulation, designed to reduce the dimensionality of a finite attribute space, both numerical and non-numerical. An illustrative example of constructing a unified block of set encapsulation is given as a solution to the problem of classifying multi-attribute objects. Practical applications of methods and technologies for multi-criteria choice of objects in a high-dimensional attribute space are presented, the main component of which is the proposed apparatus. Reducing the number of attributes allows us to simplify the solution of applied problems and meaningfully explain the results obtained.

Keywords: set encapsulation, attribute space, dimensionality reduction, aggregation tree.

DOI: 10.14357/20718594250301