INCLUSION: Return level using Kriging and covariance model from a neural network.

[ibkvictor]

Hi there,

I am working on a neural network for spatial extremes simulation. The model hopes to learn / better estimate dependent or independent extremes. I intend to add some of the implementation as part of the gstools toolset. How do I go about this?

Best,
Victor.

[LSchueler]

Hi Victor,

I'm not exactly sure what kind of an answer you are hoping for, as your question is fairly general. As this seems to be a topic which doesn't fit into any of the existing submodules like field, krige, variogram, ..., I would create a new folder in gstools/src/gstools/ and start there.
Have you thought about maybe creating your own project and using GSTools as a package from the outside? With so little information I cannot really say if your project fits into the scope of GSTools.
We have several other projects, all gathered as the GeoStat Framework.
But who knows, maybe GSTools would be a perfect match.