Hi Jose, Most kernel density estimators are straight forward to implement using the tools from the gsl statistics package. I've written a simple one for a current project using gsl functions. Here's the source<http://github.com/cboettig/structured-populations/blob/master/src/kde.c>and header<http://github.com/cboettig/structured-populations/blob/master/src/kde.h>files. The links point to my github project page in case you want to see how I'm calling the kerneldensity. Will probably expand it eventually, would like to hear what you settle on. hope this helps,
-Carl On Fri, May 7, 2010 at 7:57 AM, Jose Carlos Rubio Ballester < josecrubi...@gmail.com> wrote: > Hi, > > Does anyone know how can I train an arbitrary density distribution from > training data using a non-parametric method, like a Kernel Density > Estimator?. I see that the former "Goose" library has this feature: > > - Kernel density estimation using Epanechnikov, Biweight, Triweight, > Gaussian and Uniform kernels. > > However, I don't see anything like that in the GSL documentation. I wonder > if this algorithm is still in the library, and in that case, where can I > find information about how to use it? > > Thanks! > _______________________________________________ > Help-gsl mailing list > Help-gsl@gnu.org > http://lists.gnu.org/mailman/listinfo/help-gsl > -- Carl Boettiger Population Biology, UC Davis http://two.ucdavis.edu/~cboettig _______________________________________________ Help-gsl mailing list Help-gsl@gnu.org http://lists.gnu.org/mailman/listinfo/help-gsl
