37. évfolyam, 1991. 3. szám

Archívum

From automatic classification to higher-order morphologies.

DARÁNYI Sándor

In multivariate statistics, clustering and factoring algorithms are known to classify a set of data/cases by many variables. Such a classification is robust if different algorithms yield the same result. By a repeated classification process, robust or filtered extracts can be conceptualized as higher-order morphologies, following an astronomical metaphor. According to this, factor analytical space in observation corresponds to a universe and hitherto unseen clusters of objects/cases in it to galaxies, mapped at invisible wavelength. The disciplinary mapping of such morphs, the modelling of their evolution and navigation among them are some of the possibilities. (pp. 418-422)

Országos Széchényi Könyvtár Észrevételek