Calcualte least square cross validation bandwidth

Share:

Description

Function to estimate 2 dimensional least square cross validation for kernel density estimation

Usage

1
2
  rhrHlscv(xyz, range = c(0.1, 2), whichMin = "global",
    rescale = "unitvar", seqLength = 100)

Arguments

xyz

data.frame with three columns: x and y coordinates and of each grid cell and the number of points that are located within this cell.

range

numeric vector prortion of the reference bandwidth that are used to calculate the candidate values for least square cross validation. The default value of c(0.1, 2) will create a sequence from 0.1*href to 2*href of length seqLength.

seqLength

numeric value, the length of the sequence with candidate bandwidths.

whichMin

character, indicating if the global or local minimum should be searched

rescale

character, indicating if and if how data are rescaled. unitvar rescales x and y coordinates to unit variance, xvar rescales x and y coordinate to variance of x and none uses the raw data.

Details

Function to calcualte least square cross validation bandwidth. This implementation is based on Seaman and Powell (1996), but uses a binned algorithm. If whichMin is "global" the global minimum is returned, else the local minimum with the largest candidate bandwidth is returned.

Value

vector of length two

Author(s)

Johannes Signer

References

Seaman, D. E., & Powell, R. A. (1996). An evaluation of the accuracy of kernel density estimators for home range analysis. _Ecology, 77(7)_, 2075-2085.

Carr and Rodges

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.