Electronic Journal of Statistics

An oracle approach for interaction neighborhood estimation in random fields

Matthieu Lerasle and Daniel Y. Takahashi

Full-text: Open access

Abstract

We consider the problem of interaction neighborhood estimation from the partial observation of a finite number of realizations of a random field. We introduce a model selection rule to choose estimators of conditional probabilities among natural candidates. Our main result is an oracle inequality satisfied by the resulting estimator. We use then this selection rule in a two-step procedure to evaluate the interacting neighborhoods. The selection rule selects a small prior set of possible interacting points and a cutting step remove from this prior set the irrelevant points.

We also prove that the Ising models satisfy the assumptions of the main theorems, without restrictions on the temperature, on the structure of the interacting graph or on the range of the interactions. It provides therefore a large class of applications for our results. We give a computationally efficient procedure in these models. We finally show the practical efficiency of our approach in a simulation study.

Article information

Source
Electron. J. Statist. Volume 5 (2011), 534-571.

Dates
First available in Project Euclid: 15 June 2011

Permanent link to this document
https://projecteuclid.org/euclid.ejs/1308143122

Digital Object Identifier
doi:10.1214/11-EJS618

Mathematical Reviews number (MathSciNet)
MR2813554

Zentralblatt MATH identifier
1274.62641

Subjects
Primary: 62M40: Random fields; image analysis
Secondary: 62M45: Neural nets and related approaches

Keywords
Ising model model selection computationally efficient algorithm

Citation

Lerasle, Matthieu; Takahashi, Daniel Y. An oracle approach for interaction neighborhood estimation in random fields. Electron. J. Statist. 5 (2011), 534--571. doi:10.1214/11-EJS618. https://projecteuclid.org/euclid.ejs/1308143122.


Export citation

References

  • [1] Arlot, S. and Massart, P. (2009). Data-driven calibration of penalties for least-squares regression., Journal of Machine learning research 10 245–279.
  • [2] Baudry, J.-P., Maugis, K. and Michel, B. (2010). Slope heuristics: overview and implementation., INRIA report, available at http://hal.archives-ouvertes.fr/hal-00461639/fr/.
  • [3] Bento, J. and Montanari, A. (2009). Which graphical models are difficult to learn?, available on Arxive http://arxiv.org/pdf/0910.5761.
  • [4] Besag, L. (1993). Statistical analysis of dirty pictures., Journal of applied statistics 20 63–87.
  • [5] Birgé, L. and Massart, P. (2007). Minimal penalties for Gaussian model selection., Probab. Theory Related Fields 138 33–73.
  • [6] Bousquet, O. (2002). A Bennett concentration inequality and its application to suprema of empirical processes., C. R. Math. Acad. Sci. Paris 334 495–500.
  • [7] Bresler, G., Mossel, E. and Sly, A. (2008)., Approximation, Randomization and Combinatorial Optimization. Algorithms and Techniques Reconstruction of Markov Random Fields from Samples: Some Easy Observations and Algorithms 343-356. Springer.
  • [8] Brown, E., Kass, R. and Mitra, P. (2004). Multiple neural spike train data analysis: state-of-the-art and future challenges., Nature Neuroscience 7 456-461.
  • [9] Cross, G. and Jain, A. (1983). Markov Random field texture models., IEEE Trans. PAMI 5 25–39.
  • [10] Csiszar, I. and Talata, Z. (2006). Consistent estimation of the basic neighborhood of Markov random fields., Annals of Statistics 34 123-145.
  • [11] Galves, A., Orlandi, E. and Takahashi, D. Y. (2010). Identifying interacting pairs of sites in infinite range Ising models., Preprint, http://arxiv.org/abs/1006.0272.
  • [12] Georgii, H. (1988)., Gibbs measure and phase transitions. de Gruyter studies in mathematics 9. de Gruyter, Berlin.
  • [13] Kolmogorov, A. and Tikhomirov, V. (1963). -entropy and, ε-capacity of sets in functional spacesε. Amer.Math. Soc. Trans. 1 277–364.
  • [14] Lerasle, M. (2009). Optimal model selection in density estimation., available on Arxive http://arxiv.org/abs/0910.1654.
  • [15] Li, X., Ouyang, G., Usami, A., Ikegaya, Y. and Sik, A. (2010). Scale-free topology of the CA3 hippocampal network: a novel method to analyze functional neuronal assemblies., Biophysics Journal 98 1733-1741.
  • [16] Massart, P. (2007)., Concentration inequalities and model selection. Lecture Notes in Mathematics 1896. Springer, Berlin. Lectures from the 33rd Summer School on Probability Theory held in Saint-Flour, July 6–23, 2003, With a foreword by Jean Picard.
  • [17] Ravikumar, P., Wainwright, M. and Lafferty, J. D. (2010). High-Dimensional Ising Model Selection Using, l_1-regularized Logistic Regression. Ann. Statist. 38 1287–1319.
  • [18] Ripley, B. (1981)., Spatial Statistics. Wiley, New York.
  • [19] Schneidman, E., Berry, M., Segev, R. and Bialek, W. (2006). Weak pairwise correlations imply strongly correlated network states in a neural population., Nature 440 1007–1012.
  • [20] Takahashi, N., Sasaki, T., Matsumoto, W. and Ikegaya, Y. (2010). Circuit topology for synchronizing neurons in spontaneously active networks., Proceedings of National Academy of Science U.S.A. 107 10244-10249.
  • [21] Woods, J. (1978). Markov Image Modeling., IEEE Trans. Automat. Control 23 846–850.
  • [22] Zhou, S. (2010). Thresholded Lasso for high dimensional variable selection and statistical estimation., arXiv:1002.1583v2.