FAMILY OF GRAPH DECOMPOSITIONS AND ITS APPLICATIONS TO DATA ANALYSIS

2016. No. WP7/2016/09.

Рубчинский А. А.

A new decomposition approach to complex systems analysis is suggested. The conventional approach deals with the construction of a single, “the most correct”, decomposition of the considered system. Meanwhile the suggested approach is oriented to the construction of a family of decompositions, whose properties reveal some important meaningful features of the initial system. The expedience and applicability of the elaborated approach are illustrated by three well-known and important cases: automatic classification, political voting body and stock market. In these cases, the presented results cannot be obtained by other known methods. These examples confirm the advantages of the suggested approach.

Ключевые слова: фондовый рынок, кризис, entropy, энтропия, crisis, automatic classification, автоматическая классификация, stock market, graph decomposition, complexity index, voting political body, декомпозиция графов, показатель сложности, голосование в политических органах