Classification techniques in complex spatial databases. On the assessment of a network of world cities.

Abstract

In linking the power centers of the world-economy, a network of world cities provides the spatial outline for the reproduction of society as a capitalist world-system. An exploratory analysis of this global urban system is necessary to attain insight in its functioning, but specifications and analyses based on the use of classic data analysis techniques are hampered by the fact that they cannot assess the various sources of vagueness in this complex network of world cities. It is argued that by replacing the premises of the classic two-valued framework of conventional mathematics by a fuzzy set-theoretical approach where degrees of membership are computed rather than a mere assessment of crisp memberships in clusters, the inherent vagueness of possible classifications of world cities can be taken into account. This assertion is tested by comparing the results of some mainstream data analysis techniques (principal component analysis, crisp c-means clustering) to the results of a classification based on the premises of fuzzy set theory (fuzzy c-means clustering).