Open Rykath opened 2 years ago
One could, instead of optimizing, even integrate out hyperparameters (scale and noise). See https://towardsdatascience.com/how-to-build-a-bayesian-ridge-regression-model-with-full-hyperparameter-integration-f4ac2bdaf329?gi=d6517e1d62bf and/or ask Sascha Ranftl. The posterior will not be Gaussian then though.
Optimization of noise (
sigma_n
,sigma_p
) and of the number of basis functions (order
). Similar tooptimize
for Gaussian Processes.Good Summary: http://krasserm.github.io/2019/02/23/bayesian-linear-regression/