MATRIX SCALE
Name:
Type:
Purpose:
Description:
For some computations, such as computing a distance matrix,
it may be desirable to scale the matrix first. The scaling
may be performed over either rows or columns.
Dataplot supports the following types of scaling.
- MEAN - subtract the column mean from each column of
the matrix (or subtract the row mean from each row).
- SD - divide each column by the column standard
deviation (or divide each row by the row standard
deviation).
- ZSCORE - subtract the column mean and then divide by
the column standard deviation for each column (or
subtract the row mean and divide by the row standard
deviation for each row).
- RANGE - divide each column by the column range
(or divide each row by the row range).
The type of scaling is specified by entering the command
SET MATRIX SCALE <NONE/MEAN/SD/ZSCORE/RANGE>
Specifying NONE applies no scaling.
Syntax 1:
Syntax 2:
LET <mat2> = MATRIX COLUMN SCALE <mat1>
where <mat1> is a matrix for which the matrix scaling
is to be computed;
and where <mat2> is a matrix where the resulting
scaled matrix is saved.
This syntax scales relative to columns. The scaled matrix
can be the same as the original data matrix.
Examples:
Note:
Matrices are created with either the READ MATRIX command or the
MATRIX DEFINITION command. Enter HELP MATRIX DEFINITION and HELP
READ MATRIX for details.
Default:
Synonyms:
Related Commands:
READ MATRIX
|
= Read a matrix.
|
MATRIX COLUMN DIMENSION
|
= Dimension maximum number of
columns for Dataplot matrices.
|
MATRIX DISTANCE
|
= Compute a distance matrix.
|
Reference:
"Applied Multivariate Statistical Analysis", Third Edition,
Johnson and Wichern, Prentice-Hall, 1992.
Applications:
Implementation Date:
Program:
DIMENSION 200 COLUMNS
SKIP 25
READ IRIS.DAT SEPLENG SEPWIDTH PETLENG PETWIDTH TAG
SKIP 0
LET NTOT = SIZE SEPLENG
LET X = MATRIX DEFINITION SEPLENG NTOT 4
SET MATRIX SCALE ZSCORE
LET X = MATRIX COLUMN SCALING X
Date created: 6/5/2001
Last updated: 4/4/2003
Please email comments on this WWW page to
[email protected].
|