guiludwig/bsplinedef: Spatial Deformation via Tensor Product of B-splines

Computes coordinate transformations of the form (y1, y2) = (f1(x1, x2), f2(x1, x2)) for spatial regression, where a spatial process Y on (y1, y2) has known stationary covariance function. The functions f1 and f2 are obtained via the tensor product of B-splines, with a regularization penalty to ensure f1, f2 are injective functions. The case for Y Gaussian with general covariance function is implemented, as well as documentation for extensions to different spatial covariance functions.

install.packages("remotes")
remotes::install_github("guiludwig/bsplinedef")