A consistent model selection procedure for Markov random fields based on penalized pseudolikelihood
Chuanshu Ji and Lynne Seymour
Source: Ann. Appl. Probab. Volume 6, Number 2 (1996), 423-443.
Abstract
Motivated by applications in texture synthesis, we propose a model selection procedure for Markov random fields based on penalized pseudolikelihood. The procedure is shown to be consistent for choosing the true model, even for Gibbs random fields with phase transitions. As a by-product, rates for the restricted mean-square error and moderate deviation probabilities are derived for the maximum pseudolikelihood estimator. Some simulation results are presented for the selection procedure.
Primary Subjects: 62M40
Secondary Subjects: 62F12, 68U10
Keywords: Markov random fields; Gibbs random fields; model selection; pseudolikelihood; texture synthesis; image analysis
Full-text: Open access
Links and Identifiers
Permanent link to this document: http://projecteuclid.org/euclid.aoap/1034968138
Mathematical Reviews number (MathSciNet):
MR1398052
Digital Object Identifier: doi:10.1214/aoap/1034968138
Zentralblatt MATH identifier:
0856.62082
References
AKAIKE, H. 1974 . A new look at statistical model identification. IEEE Trans. Automat. Control 10 716 723. Z .
Mathematical Reviews (MathSciNet):
MR54:11691
AKAIKE, H. 1978 . A Bayesian analysis of the minimum AIC procedure. Ann. Inst. Statist. Math. 30 9 14.Z . BARNDORFF-NIELSEN, O. 1978 . Information and Exponential Families in Statistical Theory. Wiley, New York. Z .
BESAG, J. 1974 . Spatial interaction and the statistical analysis of lattice sy stems. J. Roy. Statist. Soc. Ser. B 6 192 236. Z .
Mathematical Reviews (MathSciNet):
MR51:9409
BROWN, L. D. 1986 . Fundamentals of Statistical Exponential Families with Applications in Statistical Decision Theory. IMS, Hay ward, CA. Z .
COMETS, F. 1992 . On consistency of a class of estimators for exponential families of Markov ´ random fields on the lattice. Ann. Statist. 20 455 468. Z .
Mathematical Reviews (MathSciNet):
MR93c:82024
CROSS, G. and JAIN, A. 1983 . Markov random field texture models. IEEE Transactions on Pattern Analy sis and Machine Intelligence 5 25 39. Z .
GEMAN, D. 1991 . Random Fields and Inverse Problems in Imaging. Lecture Notes in Math. 1427 113 193. Springer, New York. Z .
GEMAN, S. and GEMAN, D. 1984 . Stochastic relaxation, Gibbs distributions, and Bayesian restoration of images. IEEE Transactions on Pattern Analy sis and Machine Intelligence 6 721 741. Z .
Zentralblatt MATH:
0573.62030
GEMAN, S. and C. GRAFFIGNE 1986 . Markov random field image models and their applications to computer vision. Proceedings of the International Congress of Mathematicians 1496 1517. Berkeley, CA. Z .
GEORGII, H. O. 1988 . Gibbs Measures and Phase Transitions. de Gruy ter, Berlin. Z .
Mathematical Reviews (MathSciNet):
MR89k:82010
GEy ER, C. J. and THOMPSON, E. A. 1992 . Constrained Monte Carlo maximum likelihood for dependent data. J. Roy. Statist. Soc. Ser. B 54 657 699.
Mathematical Reviews (MathSciNet):
MR94f:62042
GIDAS, B. 1988 . Consistency of maximum likelihood and maximum pseudo-likelihood estimators for Gibbs distributions. In Stochastic Differential Sy stems, Stochastic Control Z . Theory and Applications W. Fleming and P.-L. Ly ons, eds. 129 146. Springer, New York. Z .
Mathematical Reviews (MathSciNet):
MR89f:62026
GIDAS, B. 1993 . Parameter estimation for Gibbs distributions from fully observed data. In Z . Markov Random Fields: Theory and Applications R. Chellappa and A. Jain, eds. 471 498. Academic Press, New York. Z .
HASSNER, M. and SKLANSKY, J. 1980 . The use of Markov random fields as models of texture. Computer Graphics and Image Processing 12 357 370. Z .
JI, C. 1990 . Sieve estimators for pair-interaction potentials and local characteristics in Gibbs random fields. Technical Report 2037, Dept. Statistics, Univ. North Carolina. Z .
KARR, A. 1991 . Statistical models and methods in image analysis: a survey. In Inference for Z . Stochastic Processes I. V. Basawa and N. U. Prabhu, eds. . Dekker, New York. Z .
Mathematical Reviews (MathSciNet):
MR92i:62188
KASHy AP, R. and CHELLAPPA, R. 1983 . Estimation and choice of neighbors in spatial interaction models of images. IEEE Trans. Inform. Theory 29 60 72. Z .
KUNSCH, H., GEMAN, S. and KEHAGIAS, A. 1995 . Hidden Markov random fields. Ann. Appl. ¨ Probab. 5 577 602. Z .
Mathematical Reviews (MathSciNet):
MR97a:60070
RIPLEY, B. D. 1981 . Spatial Statistics. Wiley, New York. Z .
Mathematical Reviews (MathSciNet):
MR83a:62217
RISSANEN, J. 1984 . Stochastic complexity. J. Roy. Statist. Soc. Ser. B 49 223 239. Z .
ROSENFELD, A. 1993 . Image modeling during the 1980's: a brief overview. In Markov Random Z . Fields: Theory and Applications R. Chellappa and A. Jain, eds. . Academic Press, New York. Z .
RUBIN, H. and SETHURAMAN, J. 1965 . Probabilities of moderate deviations. Sankhy a Ser. A 27 325 346. Z .
Zentralblatt MATH:
0178.53802
SCHWARZ, G. 1978 . Estimating the dimension of a model. Ann. Statist. 6 461 464. Z .
Zentralblatt MATH:
0379.62005
SEy MOUR, L. 1993 . Parameter estimation and model selection in image analysis using Gibbs Markov random fields. Ph.D. dissertation, Dept. Statistics, Univ. North Carolina. Z .
SEy MOUR, L. and JI, C. 1996 . Approximate Bay es model selection criteria for Markov random fields. J. Statist. Plann. Inference 51 75 97. Z .
SMITH, K. and MILLER, M. 1990 . A Bayesian approach incorporating Rissanen complexity for learning Markov random field texture models. Proceedings of the 15th International Conference on Acoustics, Speech, and Signal Processing 4 2317 2320. Z .
WOODROOFE, M. 1982 . On model selection and the arcsine laws. Ann. Statist. 10 1182 1194.
Mathematical Reviews (MathSciNet):
MR84b:62049
CHAPEL HILL, NORTH CAROLINA 27599-3260 ATHENS, GEORGIA 30602-1952 E-MAIL: cji@stat.unc.edu E-MAIL: sey mour@rolf.stat.uga.edu
The Annals of Applied Probability
