Data-driven neighborhood selection of a Gaussian field
نویسنده
چکیده
We study the non-parametric covariance estimation of a stationary Gaussian field X observed on a lattice. To tackle this issue, we have introduced a model selection procedure in a previous paper [Ver09]. This procedure amounts to selecting a neighborhood m̂ by a penalization method and estimating the covariance of X in the space of Gaussian Markov random fields (GMRFs) with neighborhood m̂. Such a strategy is shown to satisfy oracle inequalities as well as minimax adaptive properties. However, it suffers several drawbacks which make the method difficult to apply in practice. On the one hand, the penalty depends on some unknown quantities. On the other hand, the procedure is only defined for toroidal lattices. Our contribution is threefold. We propose a data-driven algorithm for tuning the penalty function. Moreover, we extend the procedure to non-toroidal lattices. Finally, we study the numerical performances of this new method on simulated examples. These simulations suggest that Gaussian Markov random field selection is often a good alternative to variogram estimation. Key-words: Gaussian field, Gaussian Markov random field, Data-driven calibration, model selection, pseudolikelihood. ∗ Laboratoire de Mathématiques UMR 8628, Université Paris-Sud, 91405 Osay † INRIA Futurs, Projet SELECT, Université Paris-Sud, 91405 Osay in ria -0 03 53 26 0, v er si on 1 15 J an 2 00 9 Sélection automatique de voisinage d’un champ gaussien Résumé : Nous étudions l’estimation non-paramétrique d’un champ gaussien stationnaire X observé sur un réseau régulier. Dans ce cadre, nous avons précédemment introduit une procédure de sélection de modèle [Ver09]. Cette procédure revient à sélectionner un voisinage m̂ grâce une technique de pénalisation puis à estimer la covariance du champ X dans l’espace des champs de Markov gaussiens de voisinage m̂. Une telle stratégie satisfait des inégalités oracles et des propriétés d’apdaptation au sens minimax. En pratique, elle présente néanmoins quelques inconvénients. D’une part, la pénalité dépend de quantités inconnues. D’autre part, la procédure est uniquement définie pour des réseaux toriques. La contribution de cet article est triple. Nous proposons un algorithme automatique pour calibrer la pénalité. De plus, nous introduisons une extension à des réseaux non-toriques. Enfin, nous étudions les performances pratiques de la procédure sur des données simulées. Ces simulations suggèrent que la sélection de champs de Markov gaussiens est souvent une bonne alternative à l’estimation de variogramme. Mots-clés : Champ gaussien, champ de Markov gaussien, calibration automatique, sélection de modèle, pseudo-vraisemblance. in ria -0 03 53 26 0, v er si on 1 15 J an 2 00 9 Neighborhood selection 3
منابع مشابه
Negative Selection Based Data Classification with Flexible Boundaries
One of the most important artificial immune algorithms is negative selection algorithm, which is an anomaly detection and pattern recognition technique; however, recent research has shown the successful application of this algorithm in data classification. Most of the negative selection methods consider deterministic boundaries to distinguish between self and non-self-spaces. In this paper, two...
متن کاملADAPTIVE ESTIMATION OF STATIONARY GAUSSIAN FIELDS BY NICOLAS VERZELEN1 INRA and SUPAGRO
We study the nonparametric covariance estimation of a stationary Gaussian field X observed on a regular lattice. In the time series setting, some procedures like AIC are proved to achieve optimal model selection among autoregressive models. However, there exists no such equivalent results of adaptivity in a spatial setting. By considering collections of Gaussian Markov random fields (GMRF) as a...
متن کاملDoes Participation in Farmer Field School Extension Program Improve Crop Yields? Evidence from Smallholder Tea Production Systems in Kenya
Agricultural Extension services are among the most important rural services in developing countries. The services are considered to be a key driver of technological change and productivity growth in agriculture. In Kenya, like in the rest of the developing economies, agricultural extension has largely been delivered through supply–driven approaches. Due to perceived low impact of agricultural e...
متن کاملEfficient Neighborhood Selection for Gaussian Graphical Models
This paper addresses the problem of neighborhood selection for Gaussian graphical models. We present two heuristic algorithms: a forward-backward greedy algorithm for general Gaussian graphical models based on mutual information test, and a threshold-based algorithm for walk summable Gaussian graphical models. Both algorithms are shown to be structurally consistent, and efficient. Numerical res...
متن کاملNovel Radial Basis Function Neural Networks based on Probabilistic Evolutionary and Gaussian Mixture Model for Satellites Optimum Selection
In this study, two novel learning algorithms have been applied on Radial Basis Function Neural Network (RBFNN) to approximate the functions with high non-linear order. The Probabilistic Evolutionary (PE) and Gaussian Mixture Model (GMM) techniques are proposed to significantly minimize the error functions. The main idea is concerning the various strategies to optimize the procedure of Gradient ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- Computational Statistics & Data Analysis
دوره 54 شماره
صفحات -
تاریخ انتشار 2010