aspcol.kernelinterpolation.get_interpolation_params

aspcol.kernelinterpolation.get_interpolation_params(kernel_func, reg_param, output_arg, data_arg, *args)

Calculates parameter vector or matrix given a kernel function for Kernel Ridge Regression. The returned parameter Z is the optimal interpolation filter from the data points to the output points. Apply filter as Z @ y, where y are the labels for data at data_arg positions

Parameters:

• data_arg (ndarray (num_data_points, data_dim))

• output_arg (ndarray (num_out_points, data_dim))

• kernel_func (function) – with calling signature kernel_func(output_arg, data_arg, *args) should return ndarray (…, num_out_points, num_data_poins)

Returns:

params

Return type:

ndarray (…, num_out_points, num_data_points)

Notes

This quantity is highly related to get_krr_params. Both are almost complete expressions of the optimal estimated function evaluated at a point, but missing one argument. This function is missing the data, and get_krr_params is missing the point at which to evaluate the estimated function.