Library: | smoother |
See also: | denbwsel denbwcrit |
Quantlet: | denrotp | |
Description: | determines a rule-of-thumb bandwidth for multivariate density estimation according to Scott. |
Usage: | hrot = denrotp(x {,K {,opt}}) | |
Input: | ||
x | n x 1 vector, the data. | |
K | string, univariate kernel function on [-1,1] or Gaussian kernel "gau". The rule-of-thumb is determined for the product kernel obtained from K. If K is not given, "qua" is used. | |
opt | string, if opt equals "robust", a robust estimate for the standard deviations of the data is used. | |
Output: | ||
hrot | scalar, rule-of-thumb bandwidth. |
library("smoother") x=read("nicfoo") hrot=denrotp(x) hrot
Contents of hrot [1,] 0.60341 [2,] 0.41691
library("smoother") x=read("nicfoo") hrot=denrotp(x,"qua","robust") hrot
Contents of hrot [1,] 0.6046 [2,] 0.44921