Quote
W. Müller, "Multivariate Statistics in Quantitative Marketing: Concept and Applications of Cluster Analysis," 2004.
Content
One of the main problems in the empirical social sciences is to capture extensive totality of objects on the basis of relevant characteristics and to divide them into groups that can be described or interpreted objectively. Such subgroups can represent both natural groupings (e.g. buyers, non-buyers of a product type) and the result of a statistical classification procedure. The object of cluster analysis is a heterogeneous set of objects of investigation (e.g. people, products, companies, regions), which are combined into subgroups (clusters, classes) that are as homogeneous as possible on the basis of object characteristics relevant to the investigation and by means of special fusion algorithms (cf. Aaker/Kumar/Day 2001, p. 566 ff.; Bacher 1996; Backhaus et. al. 2003, p. 480 ff.; Böhler 2004, p. 230 ff.; Bortz 1993, p. 522 ff.; Büschken/von Thaden 2000; Churchill/Iacobucci 2005, p. 585 ff.; Eckey/ Kosfeld/Rengers 2002; p. 203 ff.; Hammann/Erichson 2000, p. 270 ff.; Hüttner 1997, p. 319 ff.; Litz 2000; p. 384 ff.; Malhotra 1999, p. 610 ff.; Rudolf/Müller 2004, p. 151 ff.; Sudman/Blair 1998, p. 558 ff.; Voß 2004, p. 565 ff.).
Keywords
Cluster analysis
Multivariate statistics
QuantitativeMarketing