1
$\begingroup$

For traditional ridge regression, the loss function is

$loss\_function = ||A\mathbf{x}-\mathbf{b}||_2^2 + ||\Gamma\mathbf{x}||_2^2$

https://en.wikipedia.org/wiki/Tikhonov_regularization

Is there a matrix formalism to extend this such that $\Gamma\mathbf{x}$ is instead a non linear function of $\mathbf{x}$? That is

$loss\_function = ||A\mathbf{x}-\mathbf{b}||_2^2 + ||f(r,\mathbf{x})||_2^2$

Where $f(r,\mathbf{x})$ is a non linear function of $\mathbf{x}$ with $r$ being a set of tunable parameters.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.