A multigrid method for the estimation of geometric anisotropy in environmental data from sensor networks

A Review of Nonparametric Hypothesis Tests of Isotropy Properties in Spatial Data

“…Li et al (2008a) construct a test of the covariance structure for multivariate spatiotemporal data. Tests for isotropy have also been developed in the computer science literature (e.g., Molina and Feito, 2002;Chorti and Hristopulos, 2008;Spiliopoulos et al, 2011;Thon et al, 2015).…”

Inference of locally varying anisotropy fields from diverse data sources

“…There are other techniques that could be used to obtain this orientation, such as: the moment of inertia (Hassanpour, 2007), but we have found the gradient to be more stable, or; Hristopulos (2002) but this is only implemented in 2D and requires a simplex search optimization at each window that would be CPU intensive for large 3D grids, or; Spiliopoulos et al (2011) however they require differentiability of a local covariance function, no covariance function is assumed in the above workflow.…”

Stochastic Local Interaction (SLI) model: Bridging machine learning and geostatistics

“…The present version of the SLI model does not involve anisotropy. Nevertheless, anisotropy is important in cases such as the radioactivity emergency data [33]: the best performing method in SIC 2004 for this set was a general regression neural network with an anisotropic Gaussian kernel function. Similarly, in SLI it is possible to use weighted Euclidean distances or Minkowski metrics instead of the classical Euclidean distance [4].…”