The work is devoted to the problem of evaluating the proximity of multidimensional objects, the characteristics of which are measured in different scales, and the data being processed are of large dimension and contain specific text fields and omissions. Data with such specific features can not be directly processed with the classical algorithms of clustering and classification. A generalized metric is proposed in the multidimensional space of such objects, which makes it possible to build algorithms for clustering, classifications and associations based on it, using classical methods.
multidimensional objects, clusterization, classification, measurement scale, quantitative metric, categorical metric, rank metric, text metric
"Obobshchennaia metryka v zadache analyza mnohomernыkh dannыkh s raznotypnыmy pryznakamy" [Generalized metrics in the problem o f analysis of multidimensional data with different scales],
Scientific Works of Kharkiv National Air Force University,