Application of the metric data analysis method to social development indicators analysis

Abstract

The article contains a methodology for social statistics assessing. The significance of minorities (groups that differ in their attributes from the majority) has grown substantially in the modern postindustrial economy and society. In the multidimensional characteristics space distribution analysis subjects that not included in the sample can be negligible, but they may be metrically significant. For example, they can be located compactly and remotely from the main mass of subjects, i.e. have similar characteristics and significantly influence the dynamics of socio-economic systems. In addition, it is necessary to evaluate not the probability of errors in each characteristic separately, but the probability of errors in their combinations. The projection of a multidimensional space into two-dimensional or three-dimensional space, usually used to analyze such data, leads to the loss of information about compact minorities that merge into a "false majority" or the appearance of distribution. Thus, the posed task can not be reliably solved without analysis in a multidimensional space. This problem is especially relevant in states aggregated social statistics analysis for which relatively small number of subjects (countries) are represented. The aim of the article is to develop tools for the metric analysis of aggregated social statistics, which is characterized by a small number of measurements. It is developed to check the existence of stable metric and topology structures (clusters) and distribution of countries based on UN data on countries social development characteristics.