Multidimensional Classification Method in the Study of Natural and Anthropogenic Systems

Abstract

In this article the problem of build and use classification of multivariate observations; this issue has played a leading role in the knowledge of the world around us. When studying natural- anthropogenic systems have the greatest value probably-statistical simulation methods of objects, phenomena and processes. Among classification technologies to focus on situations of constructing classifications if there is no a priori information about the taxonomic structure of the observations. In the work described the rationale for the original statistical criterion Z2 and G-classification methods of multivariate observations. This used quantity conversion method dependent traits in an equivalent number of independent. G-method lets you select homogeneous classes and sub-classes of observations, evaluate differences between taxons, determine the informative signs and much more. The application of the classification shown in the case study method regularities of formation of chemical composition of mine waters in the Eastern Donbass. It is four main types of changes in the composition of mine waters: the first type has the most significant impact on environmental degradation in the region; the second and third types reflects the influence composition of groundwater chloride in the composition of mine water; the fourth type is the soda water, which may indicate the presence of the region's oil and gas fields.