9th International Science Congress (ISC-2019).  International E-publication: Publish Projects, Dissertation, Theses, Books, Souvenir, Conference Proceeding with ISBN.  International E-Bulletin: Information/News regarding: Academics and Research

A Pictorial Representation of Multivariate data and its limitation: An example from Anthropometric data

Author Affiliations

  • 1Department of Statistics, Burdwan University, Burdwan, West Bengal, INDIA
  • 2Biological Anthropological Unit, Indian Statistical Institute, Kolkata, West Bengal, INDIA

Res. J. Mathematical & Statistical Sci., Volume 2, Issue (1), Pages 6-9, January,12 (2014)

Abstract

Higher dimensional observations cannot be plotted to check dependency among its components and thus cannot be investigated whether these come from mixture population. However, a method is obtained for these two purposes in a simpler manner. Whole observations are realized by its first two principle components and from the scatter plot of these, nature of mixing and dependency may be obtained. Even scanning the pixel value of the plot, it is found that the distribution retrieved from pixel values and original distributions are different. It is due to limitation of scanning. The whole work is based on a large dimensional anthropological data set. Although, the first two principal components should be uncorrelated if it is from a unimodal distribution, but from a large data set, an impression of dependency is seen. It is shown that, it is due to mixing of distributions. Thus it is a way of identifying mixture population also, with higher dimensional observations. Also the impact over dependency of components is shown in presence of mixing.

References

  1. Pearson Karl., On lines and planes of closest fit to systems ofpoints in space, Philos. Mag., Ser., 6(2), 559-572 (1901)
  2. Hotelling H., Analysis of a complex of statistical variates intoprincipal components, Jour. Educ. Psych., 24, 417-441, 498-520 (1933)
  3. Bailey D.W., A comparison of genetic and environmentalprincipal component of morphogenesis in mice, Growth, 20,63-74 (1956)
  4. Kraus B.S. and Choi SA.C., A factorial analysis of theprenatal growth of the human skeleton, Growth, 22, 231-242(1958)
  5. Olson E.C. and Miller R.L, Morphological Integration. Xv +317. University of Chicago Press, Chicago, (1958)
  6. Guo Z., Yeu H. and Wang H., A modified PCA based on theminimum error entrophy, Proceeding of the 2004 AmericalControl Conference, Boston, Massachusetts June 30- July 2,2004 (2004)
  7. Tibshirani R., A comparison of some error estimates forneural network models" Neural Computation, 8, 152-163(1996)
  8. Seal B. and Sadhu S., Inference in a curved poissondistribution, Research journal of Mathematical and statisticalsciences, ISSN 2320-6047, 1(5), 6-16 (2013)
  9. Rao C.R., Linear Statistical Inference, Wiley, New York,(1973)
  10. Johnson R.A. and Wicheren D.W., Applied multivariatestatistical Analysis. Prentice Hall of India, New Delhi, (2001)