Ariadne
Ariadne copied to clipboard
evelinag
F# library for Gaussian process models
I am working with a rather large sub-set of data (922854 location-position pairs). When attempting to do GuassianProcess.Predict() an integer overflow occurs, which results in an index out of bounds...
Allow the scaling to be anisotropic, per feature etc. I use a kernel like this in skikit and the accuracy improves a lot.
The cookbook page link: http://mlg.eng.cam.ac.uk/duvenaud/cookbook/index.html result sin a 404. The link is at the bootom of this page: http://evelinag.com/Ariadne/ >[The Kernel Cookbook](http://mlg.eng.cam.ac.uk/duvenaud/cookbook/index.html) is a nice overview of different kernel functions...