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 pred prederr prunecv pruneseq prunetot ssr kuva

Macro: prune
Description: Prunes a smaller regression tree from a tree which has been created by the cartsplit procedure. Uses error-complexity criterion.

Usage: subcs = prune (cs, alfa)
Input:
cs list of vectors: data structure which represents a binary tree and is produced by cartsplit procedure, contains vectors cs.val, cs.vec, cs.mean, cs.ssr, cs.nelem. See cartsplit for the description of cs.
alfa real number >= 0: the tree subcs is the subtree of cs which minimizes the error-complexity criterion with the complexity parameter alfa.
Output:
subcs similar kind of object than cs, see cartsplit for the description of subcs. The tree subcs is the subtree of cs which minimizes the error-complexity criterion when the complexity parameter is alfa.

Note:

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)
cs=cartsplit(x,y,#(0,1))
cs
; let us choose a subtree when alfa=100
subcs=prune(cs,100)
subcs
Result:

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

Library: xclust
See also: cartsplit cartsplitopt cartcv leafnum maketr pred prederr 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 Klemela", 980323
(C) MD*TECH Method and Data Technologies, 17.8.2000