×

An alternative stochastic supervisor in discriminant analysis. (English) Zbl 0665.62063

Suggested is a model for a stochastic supervisor that is an alternative to the model proposed for the 2-group case, by T. Krishnan and S. C. Nandy [ibid. 20, 379-384 (1987; Zbl 0625.62043)]. The model uses a logistic-normal distribution for the classification variable and this results in a simple EM algorithm for parameter estimation. Illustrative comparisons are provided and generalizations are discussed.

MSC:

62H30 Classification and discrimination; cluster analysis (statistical aspects)

Citations:

Zbl 0625.62043
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Aitchison, J., The Statistical Analysis of Compositional Data (1986), Chapman and Hall: Chapman and Hall London · Zbl 0688.62004
[2] Aitchison, J.; Begg, C. B., Statistical diagnosis when basic cases are not classified with certainty, Biometrika, 63, 1-12 (1976) · Zbl 0336.62042
[3] Aitchison, J.; Dunsmore, I. R., Statistical Prediction Analysis (1975), Cambridge University Press · Zbl 0327.62043
[4] Dempster, A. P.; Laird, N. M.; Rubin, D. B., Maximum-likelihood, from incomplete data via the EM algorithm, J. R. Statist. Soc., B39, 1-38 (1977) · Zbl 0364.62022
[5] Kent, J. T.; Mardia, K. V., Fuzzy classification in signal processing, (Durrani, T. S.; etal., Mathematics in Signal Processing. Mathematics in Signal Processing, IMA Conf. Series No. 12 (1987), Oxford University Press), 213-223 · Zbl 0661.62055
[6] Krishnan, T.; Nandy, S. C., Discriminant analysis with a stochastic supervisor, Pattern Recognition, 20, 379-384 (1987) · Zbl 0625.62043
[7] Smith, A. F.M.; Skene, A. M.; Shaw, J. E.H.; Naylor, J. C.; Dransfield, M., The implementation of the Bayesian paradigm, Communs Statist. theor. meth., A14, 1079-1102 (1985) · Zbl 0582.62025
[8] Titterington, D. M.; Smith, A. F.M.; Makov, U. E., Statistical Analysis of Finite Mixture Distributions (1985), Wiley: Wiley New York · Zbl 0646.62013
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.