Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library: smoother
See also: denest

Macro: denestp
Description: estimates a p-dimensional density by kernel density estimation. The computation uses WARPing.

Usage: fh = denestp(x {,h {,K} {,d} })
Input:
x n x p matrix, the data.
h scalar or p x 1 vector, bandwidth. If not given, the rule of thumb bandwidth computed by denrotp is used (Scott's rule of thumb).
K string, kernel function on [-1,1]^p. If not given, the product Quartic kernel "qua" is used.
Output:
d
fh m x (p+1) matrix, the first p columns constitute a grid and the last column contains the density estimate on that grid.

Example:
library("smoother")                                       
library("plot")
;
x  = read("geyser")     ; read data         
fh = denestp(x)         ; estimate density
;
fh = setmask(fh,"surface")
plot(fh)                ; graph density estimate
setgopt(plotdisplay,1,1,"title","ROTATE!")
Result:
The kernel density estimate for the Geyser data is    
computed using the Quartic kernel and bandwidth
according to Scott's rule of thumb (default).
The display shows the surface of the resulting 
function.
Example:
library("smoother")                                       
library("plot")
;
x  = read("bank2.dat")            ; read data                    
x  = x[,4:6]                      ; columns 4 to 6
d  = (max(x)-min(x))'./7          ; large binwidth!         
fh = denestp(x,1.5,"qua",d)       ; estimate density
;
c  = (max(fh[,4])-min(fh[,4])).*(1:4)./5 ; levels
cfh= grcontour3(fh,c,1:4)                ; contours
plot(cfh)                         ; graph contours
setgopt(plotdisplay,1,1,"title","ROTATE!")
Result:
The kernel density estimate for the last three 
variables of the Swiss banknote data is computed 
using the Quartic kernel and bandwidth h=1.5.
The display shows a contour plot of the resulting 
function.

Library: smoother
See also: denest

Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Author: Sigbert Klinke, 950225; Lijian Yang, 960504; Marlene Mueller, 990413
(C) MD*TECH Method and Data Technologies, 28.6.1999