Advertisement

Multivariate analysis methods: Background and example

  • Fionn Murtagh
VII Methods and Tools
Part of the Lecture Notes in Physics book series (LNP, volume 310)

Abstract

Multivariate statistical methods deal with the inherently very difficult problem of detecting patterns in data. These patterns can take many forms — natural groups, inherent dimensionality, correlations, dependencies, and so on. Often, therefore, different methods bring different features of the data to light.

Following a brief overview of some prominent multivariate methods, we illustrate their use on IRAS data. We indicate how different multivariate methods can be “chained together” to yield powerful tools for uncovering structure in data.

References

  1. Adorf, H.-M., Meurs, E.J.A., 1988. These proceedings, p. 315.Google Scholar
  2. Barrow, J.D., Bhavsar, S.P., Sonoda, D.H., 1985. Mon. Not. R. astr. Soc., 216, 17.Google Scholar
  3. Meurs, E.J.A., Adorf, H.-M., Harmon, R.T., 1988. In Astronomy from Large Databases, eds. Murtagh, F., Heck, A., European Southern Observatory, Garching, p. 49.Google Scholar
  4. The MIDAS Users Guide, 1985 (and subsequent versions). Chapter 13, Multivariate Statistical Methods, European Southern Observatory, Garching.Google Scholar
  5. Murtagh, F., 1985. Multidimensional Clustering Algorithms, Physica-Verlag, Heidelberg and New York.Google Scholar
  6. Murtagh, F., Heck, A., 1987a. Astr. Astrophys. Suppl., 68, 113. (copy of bibliography available from author).Google Scholar
  7. Murtagh, F., Heck, A. 1987b. Multivariate Data Analysis, Kluwer Academic, Dordrecht.Google Scholar
  8. Rohlf, F.J., 1978. Information Processing Letters, 7, 44.CrossRefGoogle Scholar
  9. Schulman, L.S., Seiden, P.E., 1986. Science, 233, 425. *** DIRECT SUPPORT *** A3418241 00009Google Scholar

Copyright information

© Springer-Verlag 1988

Authors and Affiliations

  • Fionn Murtagh
    • 1
  1. 1.Space Telescope - European Coordinating FacilityEuropean Southern ObservatoryGarching bei MünchenGermany

Personalised recommendations