This file contains an implmentation of kernel density estimation using multidimensional version of the original fast Gauss transform for a linkable library component. More...
Go to the source code of this file.
Classes | |
class | FGTKde |
A computation class for FGT based kernel density estimation. More... |
This file contains an implmentation of kernel density estimation using multidimensional version of the original fast Gauss transform for a linkable library component.
This algorithm by design supports only the Gaussian kernel with the fixed-bandwidth. The optimal bandwidth cross-validation routine is not provided in this library.
For more details on mathematical derivations, please take a look at the following paper:
Article{ggstrain, Author = "L. Greengard and J. Strain", Title = "{The Fast Gauss Transform}", Journal = "SIAM Journal of Scientific and Statistical Computing", Volume = "12(1)", Year = "1991", Pages = "79-94" }
TO-DO: Extend this code to nonuniform weights, and also replace all series expansion calls with those in series expansion library.
Definition in file fgt_kde.h.