Consider a positive definite kernel $K:\mathbb N\times \mathbb N\rightarrow \mathbb R$. Denote the unique RKHS associated with $K$ by $\mathcal H_K$. The RKHS $\mathcal H_K$ consists of \begin{align} \mathcal H_0= \left\{\sum_{t'=1}^n a_{t'}K(\cdot,t') \mid n\in\mathbb N, a_{t'}\in\mathbb R \right\} \end{align} and its completion. The completion of $\mathcal H_0$ is the set of functions which are pointwise limits of Cauchy sequences in $\mathcal H_0$ with the inner product defined by \begin{align} \langle h_a,h_b \rangle_{\mathcal H_K} = \sum_{t=1}^n\sum_{t'=1}^m a_{t}b_{t'}K(t,t') \end{align} where $h_a=\sum_{t=1}^n a_{t}K(t,\cdot)$ and $h_b=\sum_{t'=1}^m b_{t'}K(\cdot,t')$ with $a_{t},b_{t'}\in\mathbb R$.

My question is:

If $\mathcal H_K\subset \ell^1$, is $\{f_n\}$ with $f_n = \sum_{t=1}^n a_{t}K(\cdot,t)$ and $\{a_t\}\in\ell^\infty$ a Cauchy sequence in $\mathcal H_0$?

This problem may be a basic one (?) but has bothered me for a while. Any ideas and suggestions will be great. Thank you.

I have read some papers and books regarding RKHS, e.g.,

A. Berlinet and C. Thomas-Agnan, Reproducing kernel Hilbert spaces in probability and statistics. Kluwer Academic Boston, 2004.

However, I did not find relevant discussions.


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.