[R] KernSmooth: bkde and dpik bandwidth questions

[Prof Brian Ripley]

On Wed, 2 Jan 2008, Rex Carazo-Zapetis wrote:

> Hi,
>
>   I have two separate questions relating to the KernSmooth package.  I am
> using the dpik function from the KernSmooth package and receive the error
>
> Warning message:
>    In kappam * Gcounts :
>     longer object length is not a multiple of shorter object length

>From the footer of every R-help message:

> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

We have no idea how to reproduce this.

>
>   I saw an earlier post , but the issue was using the bkde fxn and the
> person appeared to be using too small of a bandwidth.  However,
> dpik returns a bandwidth and therefore I am not sure why I am getting this
> error.  If I use dpik and specify scalest to be "stdev" then the fxn works.
>
>
>  Secondly,  I am using the bkde function and when I put in a vector of data
> the function computes some bandwidth to use as it then proceeds to return x
> and y vectors.  I do not see in  the documentation the process by which this
> bandwidth is computed.  It does not appear to match the bandwidth (or even
> be remotely close, it is an order of magnitude different) when using the
> dpik function (with scalest = "stdev").  If someone could tell me how it
> calculates the bandwidth which the function uses I would appreciate it.

This is support software for a book, referenced on the help page.  Have 
you consulted it?  A quick read of the code of bkde will show you which of 
the methods it is.

> Thank you in advance.
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595