haku: @indexterm cluster analysis / yhteensä: 167
viite: 16 / 167
Tekijä: | Franke, N. Reisinger, H. Hoppe, D. |
Otsikko: | Remaining within-cluster heterogeneity: A meta-analysis of the "dark side" of clustering methods |
Lehti: | Journal of Marketing Management
2009 : APR, VOL. 25:3-4, p. 273-293 |
Asiasana: | research methodology cluster analysis simulation marketing management |
Kieli: | eng |
Tiivistelmä: | Based on a meta-analysis of articles using clustering methods, this study finds little attention having paid to remaining within-cluster heterogeneity (hereafter as: w-clst-htgy). Average values are found to be relatively high. In order to address this potentially problematic "dark side" of cluster analysis (here as: clst-an.), provided are two coefficients as standard information in any clst-an. findings: i. a goodness-of-fit measure and ii. a measure relating explained variation (here as: exp-var.) of analysed empirical data to exp-var. of simulated random data. The second coefficient is referred to as the Index of Clustering Appropriateness (ICA). A classification scheme representing acceptable levels of remaining w-clst-htgy. is developed. |
SCIMA