Based on previous discussion with @sathvikbhagavan , derivatives for positive definite kernels (Kriging, Squared exponential functions, etc. is of interest.)
Kernels in the package include:
Radials
Inverse distance
Wenland
Kriging
Not in the package
Matern kernels
Wavelet features
Weighted random features (different from reservoir computing features)
Derivatives of order I, II, III and IV are required.
Since kernels are explicitly known and symbolics is not part of the package yet, separate dispatch functions identified by a Differential operator struct for the following kernels will be built.
Based on previous discussion with @sathvikbhagavan , derivatives for positive definite kernels (Kriging, Squared exponential functions, etc. is of interest.)
Kernels in the package include:
Not in the package
Derivatives of order I, II, III and IV are required. Since kernels are explicitly known and symbolics is not part of the package yet, separate dispatch functions identified by a Differential operator struct for the following kernels will be built.
To do so, the following is necessary: