Σύγκριση μεθόδων

	Επι-διαστατική Κλιμάκωση (MDS)	Ανάλυση Λανθανουσών Κλάσεων (LCA)
Πεδίο	Στατιστική	Στατιστική
Οικογένεια	Latent structure	Latent structure
Έτος προέλευσης	1952–1964	1950s–1968
Δημιουργός	Warren S. Torgerson (metric MDS, 1952); Joseph B. Kruskal (non-metric MDS, 1964)	Paul F. Lazarsfeld
Τύπος	Dimensionality reduction / visualization	Latent variable / person-centered classification
Θεμελιώδης πηγή	Kruskal, J. B. (1964). Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis. Psychometrika, 29(1), 1–27.	Goodman, L. A. (1974). Exploratory latent structure analysis using both identifiable and unidentifiable models. Biometrika, 61(2), 215–231.
Εναλλακτικές ονομασίες	MDS, metric MDS, non-metric MDS, proximity scaling	LCA, latent class model, latent categorical analysis, finite mixture of multinomials
Συναφείς	5	6
Σύνοψη	Multidimensional scaling maps objects described only by pairwise similarities or dissimilarities into a low-dimensional geometric space so that distances in that space reflect the original proximity structure as faithfully as possible. It is widely used to visualize the hidden structure of psychological, social, and behavioral data.	Latent class analysis identifies unobserved subgroups — latent classes — within a population by finding patterns of responses across a set of categorical observed indicators. It is the categorical-variable counterpart of cluster analysis, but grounded in an explicit probabilistic model, and is widely used in social, health, and behavioral sciences to discover typologies in survey or diagnostic data.