Let's suppose we have a kernel function

$k(x,x')=10 $

In order to prove that this a valid kernel function there are generally two conditions

  • It is symmetric
  • There exists a map $\varphi:R^d \rightarrow H$ called kernel feature map into some high dimensional feature space $H$ such that $\forall x,x' \ in \ R^d :k(x,x') = \ <\varphi(x),\varphi(x')> $.

How to formally approach to prove these two conditions for this kernel function?


If I get you correctly, you simply want $k(x, x')$ to be the constant function $10$? In that case

  • It is symmetric because for all $x, x'$: $k(x, x')=10=k(x',x)$
  • You can take $H = \mathbb{R}$ and $\varphi(x) = \sqrt{10}$. Then for any $x, x'$: $\langle\varphi(x),\varphi(x')\rangle = \sqrt{10}^2 = 10 = k(x,x')$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.