Gaussian likelihood inference on data from trans-Gaussian random fields with Matérn covariance function
KAUST DepartmentApplied Mathematics and Computational Science Program
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
Permanent link to this recordhttp://hdl.handle.net/10754/625682
MetadataShow full item record
AbstractGaussian likelihood inference has been studied and used extensively in both statistical theory and applications due to its simplicity. However, in practice, the assumption of Gaussianity is rarely met in the analysis of spatial data. In this paper, we study the effect of non-Gaussianity on Gaussian likelihood inference for the parameters of the Matérn covariance model. By using Monte Carlo simulations, we generate spatial data from a Tukey g-and-h random field, a flexible trans-Gaussian random field, with the Matérn covariance function, where g controls skewness and h controls tail heaviness. We use maximum likelihood based on the multivariate Gaussian distribution to estimate the parameters of the Matérn covariance function. We illustrate the effects of non-Gaussianity of the data on the estimated covariance function by means of functional boxplots. Thanks to our tailored simulation design, a comparison of the maximum likelihood estimator under both the increasing and fixed domain asymptotics for spatial data is performed. We find that the maximum likelihood estimator based on Gaussian likelihood is overall satisfying and preferable than the non-distribution-based weighted least squares estimator for data from the Tukey g-and-h random field. We also present the result for Gaussian kriging based on Matérn covariance estimates with data from the Tukey g-and-h random field and observe an overall satisfactory performance.
CitationYan Y, Genton MG (2017) Gaussian likelihood inference on data from trans-Gaussian random fields with Matérn covariance function. Environmetrics: e2458. Available: http://dx.doi.org/10.1002/env.2458.
SponsorsThis research was supported by the King Abdullah University of Science and Technology (KAUST).
CollectionsArticles; Applied Mathematics and Computational Science Program; Applied Mathematics and Computational Science Program; Statistics Program; Statistics Program; Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division; Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
Showing items related by title, author, creator and subject.
Advances in Gaussian random field generation: a reviewLiu, Yang; Li, Jingfa; Sun, Shuyu; Yu, Bo (Computational Geosciences, Springer Science and Business Media LLC, 2019-08-05) [Article]Gaussian (normal) distribution is a basic continuous probability distribution in statistics, it plays a substantial role in scientific and engineering problems that related to stochastic phenomena. This paper aims to review state-of-the-art of Gaussian random field generation methods, their applications in scientific and engineering issues of interest, and open-source software/packages for Gaussian random field generation. To this end, first, we briefly introduce basic mathematical concepts and theories in the Gaussian random field, then seven commonly used Gaussian random field generation methods are systematically presented. The basic idea, mathematical framework of each generation method are introduced in detail and comparisons of these methods are summarized. Then, representative applications of the Gaussian random field in various areas, especially of engineering interest in recent two decades, are reviewed. For readers’ convenience, four representative example codes are provided, and several relevant up-to-date open-source software and packages that freely available from the Internet are introduced.
On the Distribution of Indefinite Quadratic Forms in Gaussian Random VariablesAl-Naffouri, Tareq Y.; Moinuddin, Muhammed; Ajeeb, Nizar; Hassibi, Babak; Moustakas, Aris L. (IEEE Transactions on Communications, Institute of Electrical and Electronics Engineers (IEEE), 2015-10-30) [Article]© 2015 IEEE. In this work, we propose a unified approach to evaluating the CDF and PDF of indefinite quadratic forms in Gaussian random variables. Such a quantity appears in many applications in communications, signal processing, information theory, and adaptive filtering. For example, this quantity appears in the mean-square-error (MSE) analysis of the normalized least-meansquare (NLMS) adaptive algorithm, and SINR associated with each beam in beam forming applications. The trick of the proposed approach is to replace inequalities that appear in the CDF calculation with unit step functions and to use complex integral representation of the the unit step function. Complex integration allows us then to evaluate the CDF in closed form for the zero mean case and as a single dimensional integral for the non-zero mean case. Utilizing the saddle point technique allows us to closely approximate such integrals in non zero mean case. We demonstrate how our approach can be extended to other scenarios such as the joint distribution of quadratic forms and ratios of such forms, and to characterize quadratic forms in isotropic distributed random variables.We also evaluate the outage probability in multiuser beamforming using our approach to provide an application of indefinite forms in communications.
Exact Symbol Error Probability of Square M-QAM Signaling over Generalized Fading Channels subject to Additive Generalized Gaussian NoiseSoury, Hamza; Alouini, Mohamed-Slim; Yilmaz, Ferkan (2013 IEEE International Symposium on Information Theory, Institute of Electrical and Electronics Engineers (IEEE), 2013-07) [Conference Paper]This paper considers the average symbol error probability of square Quadrature Amplitude Modulation (QAM) coherent signaling over flat fading channels subject to additive generalized Gaussian noise. More specifically, a generic closedform expression in terms of the Fox H function and the bivariate Fox H function is offered for the extended generalized-K fading case. Simplifications for some special fading distributions such as generalized-K fading, Nakagami-m fading, and Rayleigh fading and special additive noise distributions such as Gaussian and Laplacian noise are then presented. Finally, the mathematical formalism is illustrated by some numerical examples verified by computer based simulations for a variety of fading and additive noise parameters.