[Development of multidimensional scaling methods in biomedical research].
Analyzed were different approaches to the construction of multidimensional scales to be used in comparison of experimental biomedical data. Pinpointed were weak points of the existing models. Multidimensional scaling algorithms can be adapted by substituting the notion of "measure of nearness" between characteristics studied with the dispersion analysis which is the most valuable part of the multidimensional methods--orthogonal scaling (from one up to three scales). Integral criterion D2 1,2 allows evaluation of both individual and typological reactions of the body systems to various factors.