You've tried loess?
On 07/02/2008, charr <[EMAIL PROTECTED]> wrote:
>
> Hi,
> i would like to do a local linear regression but with a user-defined
> function f(x,y) as kernel.
> (not the typical Gaussian or Epanechnikov kernel function but some similar
> function).
> More, would it be possible to
Hi,
i would like to do a local linear regression but with a user-defined
function f(x,y) as kernel.
(not the typical Gaussian or Epanechnikov kernel function but some similar
function).
More, would it be possible to do a local linear regression by adding weights
to the (x,y) data points?
Many than
2 matches
Mail list logo
