[Validity of epidemiological classification: construction of a discriminant criterion]. / Dostovernost' épidemiologicheskikh klassifikatsii: postroenie diskriminantnogo kriteriia.

A new criterion has been proposed to assess the validity of group division differences obtained by preliminary cluster analysis. The method was used to provide evidence for the validity of differences between the groups as to epidemiological classifications. The paper shows it expedient to use the method in the epidemiological analysis.