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: xclust
See also: cartsplit cartsplitopt cartcv leafnum maketr prederr prune prunecv pruneseq prunetot ssr kuva

Macro: pred
Description: Given a regression tree m (output of cartsplit) and a vector x, returns m(x), that is, the predicton of the tree on that point.

Usage: enn = pred (tr, x, type)
Input:
tr list of vectors: data structure which represents a binary tree and is produced by cartsplit procedure, contains vectors tr.val, tr.vec, tr.mean, tr.ssr, tr.nelem. See cartsplit for the description of tr.
x p x 1 vector: represents a point in the sample space at which the prediction of the regression tree will be given.
type p x 1 vector: contains the types of the original regression variables, 1 means that the corresponding variable is continuous and 0 that it is categorical. This vector should be similar to the vector which was given originally as an input to the cartsplit which produced the regression tree tr.
Output:
enn real number: prediction of the regression tree at x, that is, the value of the regression function at x.

Example:
; loads the library xclust
library ("xclust")
;let us generate a tree by cartsplit procedure
x1=#(0,0,0,0,1,1,1,1,1,2)
x2=#(0,0,0,0,0,0,0,1,1,1)
x=x1~x2
y=#(0,0,0,0,100,100,100,120,120,120)
tr=cartsplit(x,y,#(0,1))
tr
enn1=pred(tr,#(0,1),#(0,1))
enn1
enn2=pred(tr,#(1,-1),#(0,1))
enn2
enn3=pred(tr,#(1,1),#(0,1))
enn3
Result:

Content of object tr.val.split0
[1,] 0 
[2,] 1,2 
Content of object tr.val.split1
[1,] NaN 
Content of object tr.val.split2
[1,] 0 
Content of object tr.val.split3
[1,] NaN 
Content of object tr.val.split4
[1,] NaN 
Content of object tr.vec
[1,] 1.000000 
[2,] NaN 
[3,] 2.000000 
[4,] NaN 
[5,] NaN 
Content of object tr.mean
[1,] 66.000000 
[2,] 0.000000 
[3,] 110.000000 
[4,] 100.000000 
[5,] 120.000000 
Content of object tr.var
[1,] 29640.000000 
[2,] 0.000000 
[3,] 600.000000 
[4,] 0.000000 
[5,] 0.000000 
Content of object tr.nelem
[1,] 10.000000 
[2,] 4.000000 
[3,] 6.000000 
[4,] 3.000000 
[5,] 3.000000 
Content of object enn1
[1,] 0.000000
Content of object enn2
[1,] 100.000000
Content of object enn3
[1,] 120.000000

Library: xclust
See also: cartsplit cartsplitopt cartcv leafnum maketr prederr prune prunecv pruneseq prunetot ssr kuva

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: Jussi Klemelae, 980323
(C) MD*TECH Method and Data Technologies, 28.6.1999