Markov Chain Random Fields for Estimation of Categorical Variables
- 370 Downloads
Multi-dimensional Markov chain conditional simulation (or interpolation) models have potential for predicting and simulating categorical variables more accurately from sample data because they can incorporate interclass relationships. This paper introduces a Markov chain random field (MCRF) theory for building one to multi-dimensional Markov chain models for conditional simulation (or interpolation). A MCRF is defined as a single spatial Markov chain that moves (or jumps) in a space, with its conditional probability distribution at each location entirely depending on its nearest known neighbors in different directions. A general solution for conditional probability distribution of a random variable in a MCRF is derived explicitly based on the Bayes’ theorem and conditional independence assumption. One to multi-dimensional Markov chain models for prediction and conditional simulation of categorical variables can be drawn from the general solution and MCRF-based multi-dimensional Markov chain models are nonlinear.
KeywordsMulti-dimensional Markov chain Markov random field Conditional simulation Interclass relationship Nonlinear Conditional independence
Unable to display preview. Download preview PDF.
- Besag, J., 1974, Spatial interaction and the statistical analysis of lattice systems (with discussion): J. Roy. Stat. Soc. Ser. B, v. 36, no. 2, p. 192–236. Google Scholar
- Besag, J., 1986, On the statistical analysis of dirty pictures (with discussions): J. Roy. Stat. Soc. Ser. B, v. 48, no. 3, p. 259–302. Google Scholar
- Idier, J., Goussard, Y., Ridolfi, A., 2001, Unsupervised image segmentation using a telegraph parameterization of Pickard random fields, in Moore, M., ed., Spatial Statistics. Methodological Aspects and Some Applications, Lecture Notes in Statistics, vol. 159: Springer, New York, p. 115–140. Google Scholar
- Li, W., 2006, Transiogram: a spatial relationship measure for categorical data: Int. J. Geogr. Inf. Sci., v. 20, no. 6, p. 693–699. Google Scholar
- Li, W., Zhang, C., 2005, Application of transiograms to Markov chain simulation and spatial uncertainty assessment of land-cover classes: GISci. Remote Sens., v. 42, no. 4, p. 297–319. Google Scholar
- Lin, C., Harbaugh, J.W., 1984, Graphic display of two- and three-dimensional Markov computer models in geology: Van Nostrand Reinhold Company, New York, p. 180. Google Scholar
- Qian, W., Titterington, D.M., 1991, Multidimensional Markov chain models for image textures: J. Roy. Stat. Soc. Ser. B, v. 53, no. 3, p. 661–674. Google Scholar
- Rosholm, A., 1997, Statistical methods for segmentation and classification of images: PhD dissertation, Technical University of Denmark, Lyngby, p. 187. Google Scholar
- Ripley, B.D., 1990, Gibbsian interaction models, in Griffith, D.A., ed., Spatial statistics: Past, present, and future: Institute of Mathematical Geography, Syracuse University, p. 3–25. Google Scholar
- Sharp, W.E., Turner, B.F., 1999, The extension of a unilateral first-order autoregressive process from a square net to an isometric lattice using the herringbone method, in Lippard, S.J., Nass, A., Sinding-Larsen, R., eds., Proceedings of IMAG’99, 6–11 August 1999, Trondheim, p. 255–259. Google Scholar
- Zhang, C., Li, W., 2005, Markov chain modeling of multinomial land-cover classes: GISci. Remote Sens., v. 42, no. 1, p. 1–18. Google Scholar