I am trying to do the following: f = K1((x1, y1), (x1, y1)*) + K2((x2, y2), (x2, y2)*) + x4 * K3((x1, y1), (x1, y1)*) where K1, K2, and K3 are kernels that use two dimensions each (x and y). The third kernel (K3) needs to be scaled by x3. I thought I would perform this by K4(x4) * K3. However, I …

Gaussian Processes - combining constant features with covariance

arkottke December 1, 2017, 10:20pm 5

Yeah, I think that __init__() would just have to be modified to permit passing the input_dim and active_dims arguments.

Topic		Replies	Views
Gaussian Process regression with categorical features Questions	1	4946	June 13, 2019
Scaling covariance function in Gaussian Processes Questions	3	760	October 28, 2019
Multiple (uncertain) function observations of the same Gaussian process Questions	15	6568	October 16, 2017
ScaledCov: Implementation issue of simple scaled cov function Questions	2	471	October 3, 2018
Multidimensional input using Gaussian Process Questions	6	4028	June 28, 2017