Field generation and kriging using circulant covariance matrix and Tensor Train (TT) approximations for the paper "Kriging in Tensor Train data format" by Sergey Dolgov, Alexander Litvinenko and Dishi Liu [arXiv:1904.09668].
This package consists of two drivers:
- driver_generate_field.m: generates a random field given a circulant covariance matrix, using superposition of a Cartesian product of iid normal vectors.
- driver_cov_fft.m: interpolation of a set of samples onto a coarse Cartesian grid and kriging onto a fine grid via circulant covariance matrices.
TT algorithms use TT-Toolbox (https://github.com/oseledets/TT-Toolbox). It can be downloaded automatically using the check_tt_install script.