One-Mode Three-Way Analysis Based on Result of One-Mode Two-Way Analysis

Conference paper
Part of the Studies in Classification, Data Analysis, and Knowledge Organization book series (STUDIES CLASS)

Abstract

Several analysis models for proximity data have been introduced. While most of them were for one-mode two-way data, some analysis models which are able to analyze one-mode three-way data have been suggested in recent years. One-mode three-way multidimensional scaling and overlapping cluster analysis models were suggested. Furthermore, several studies were done for comparison between the results of analyses by one-mode two-way and by one-mode three-way data, where both of them are generated from the same source of data. In the present study, the authors suggest the analysis of one-mode three-way data based on one-mode two-way analysis for overlapping clusters. To evaluate the necessity of one-mode three-way analysis, firstly one-mode three-way data are reconstructed from the clusters and weights obtained by one-mode two-way overlapping cluster analysis. Secondly the reconstructed one-mode three-way data are subtracted from original one-mode three-way data. The subtracted data were analyzed by one-mode three-way overlapping cluster analysis model. The result of analysis discloses the components of proximities which can be expressed by one-mode three-way analysis but not by one-mode two-way analysis.

References

  1. Arabie, P., & Carroll, J. D. (1980). Mapclus: A mathematical programming approach to fitting the ADCLUS model. Psychometrika, 45, 211–235.MATHCrossRefGoogle Scholar
  2. De Rooij, M., & Heiser, W. J. (2000). Triadic distance models for the analysis of asymmetric three-way proximity data. British Journal of Mathematical and Statistical Psychology, 53, 99–119.MathSciNetCrossRefGoogle Scholar
  3. Nakayama, A. (2005). A multidimensional scaling model for three-way data analysis. Behaviormetrika, 32, 95–110.MathSciNetMATHCrossRefGoogle Scholar
  4. Shepard, R. N., & Arabie, P. (1979). Additive clustering: Representation of similarities as combinations of discrete overlapping properties. Psychological Review, 86, 87–123.CrossRefGoogle Scholar
  5. Yokoyama, S., Nakayama, A., & Okada, A. (2009). One-mode three-way overlapping cluster analysis. Computational Statistics, 24, 165–179.MathSciNetMATHCrossRefGoogle Scholar
  6. Yokoyama, S., Nakayama, A., & Okada, A. (2010). An application of one-mode three-way overlapping cluster analysis. In L. J. Hermann & W. Claus (Eds.), Classification as a tool for research (pp. 193–200). Berlin: Springer.CrossRefGoogle Scholar
  7. Yokoyama, S., & Okada, A. (2010a). Consideration about the necessity of one-mode three-way data analysis. In Proceedings of the 26th meeting of the Japanese classification society (in Japanese), Fukuoka (pp. 13–14).Google Scholar
  8. Yokoyama, S., & Okada, A. (2010b). A study on the necessity of analysis in one-mode proximity data. Proceedings of the 34th annual conference of the German classification society, Karlsruhe (p. 197).Google Scholar
  9. Yokoyama, S., & Okada, A. (2010c): External analysis of overlapping cluster analysis. Proceedings of the 28th meeting of the Japanese classification society (in Japanese), Tokyo (pp. 13–14).Google Scholar
  10. Yokoyama, S., & Okada, A. (2011a): External analysis of POS data. Proceedings of the 29th Meeting of the Japanese Classification Society (in Japanese), Kyoto (pp. 11–12).Google Scholar
  11. Yokoyama, S., & Okada, A. (2011b). A study on the necessity of one-mode three-way analysis. The Teikyo University Economic Review, 45, 143–150.Google Scholar

Copyright information

© Springer International Publishing Switzerland 2013

Authors and Affiliations

  1. 1.Department of Business AdministrationFaculty of Economics, Teikyo UniversityTokyoJapan
  2. 2.Graduate School of Management and Information SciencesTama UniversityTama, TokyoJapan

Personalised recommendations