Abstract
This paper focusses on a simulation comparison of the Adaptive Bayesian Classification (ABC) and ICM (iterated Conditional Mode) spatial classification procedures for image reconstruction. The ABC procedure is described, since it is less well-known. The same geometrical configuration of populations is used for several examples involving normal populations with increasingly different variances. It is found that for the cases studied, ABC has considerably smaller variances than ICM in percentage of correct classifications obtained. Moreover, ABC has about the same distribution of percentage of correct classifications obtained whether we examine results for each population separately, or whether we examine the overall result. The comparison is made using for ICM the stated ICM initial classification, which is a “maximum likelihood classifier”; and using for ABC, the ABC initial classification, which is a predictive, contextual Bayesian classification procedure.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Besag, J. (1986). “On the Statistical Analysis of Dirty Pictures”, Jour. Royal Statist. Soc. (B), 48, 3, 259–302.
De Finetti, B. (1937). “Foresight: Its Logical Laws, Its Subjective Sources”, in Annales de l’Institut Henri Poincare, Vol. 7, published by Gauthier-Villars; translated and reproduced in Studies in Subjective Probability, edited by Henry Kyburg and Howard Smokier, Krieger Publishing Co., Second Edition, 1980, 53–118.
Elphinstone, C. D., Lanergan, A. T., Fatti, L. P. and D. M. Hawkins (1985). “An Empirical Investigation into the Application of Some Statistical Techniques to the Classification of Remotely Sensed Data.” CSIRO Special Report SWISK 40, Pretoria, South Africa.
Geisser, S. (1964). “Posterior Odds for Multivariate Normal Classifications”, JRSS(B), 26, 69–76.
Geisser, S. (1966). “Predictive Discrimination”. Multivariate Analysis, ed. P. R. Krishnaiah. New York: Academic Press, 149–163.
Klein, R. and S. J. Press (1988). “The ABC (Adaptive Bayesian Classification) of Remotely Sensed Data”, Technical Report No. 172, Department of Statistics, University of California, Riverside.
Klein, R. and S. J. Press (1989). “Contextual Bayesian Classification of Remotely Sensed Data”, Communications in Statistics — Theory and Methods, 18(9), 3177–3202.
Klein, R. and S. J. Press (1990a). “Bayesian Contextual Classification with Neighbors Correlated with Training Data”, in Bayesian and Likelihood Methods in Statistics and Econometrics: Essays in Honor of George A. Bernard, (eds.) S. Geisser, J. Hodges, S. J. Press, A. Zellner, New York: North Holland Pub. Co., 337–355.
Klein, R. and S. J. Press (1990b). “Bayesian Classification of Remotely Sensed Data when Training Data is Part of the Scene”. Brasileira De Probabilidade E Estatistica, in English, 4, No. 1.
Mardia, K. V. (1984). “Spatial Discrimination and Classification Maps”, Communications in Statistics, Theory and Methods, 13, No. 18, 2181–2197.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1992 Springer-Verlag New York, Inc.
About this paper
Cite this paper
Klein, R., Press, S.J. (1992). Simulation Comparison of Methods for Bayesian Contextual Classification of Remotely Sensed Data. In: Goel, P.K., Iyengar, N.S. (eds) Bayesian Analysis in Statistics and Econometrics. Lecture Notes in Statistics, vol 75. Springer, New York, NY. https://doi.org/10.1007/978-1-4612-2944-5_9
Download citation
DOI: https://doi.org/10.1007/978-1-4612-2944-5_9
Publisher Name: Springer, New York, NY
Print ISBN: 978-0-387-97863-5
Online ISBN: 978-1-4612-2944-5
eBook Packages: Springer Book Archive