Help for EIGENVEC
PURPOSE:
EIGENVEC will compute the principal component transformation
matrix for up to 32 input channels. The covariance matrix, the transformation
matrix of eigen-vectors, and the eigen-values are printed. This program
should normally be run via the procedure EIGEN.
OPERATION:
The measure of inter-dimensional correlation in the multi-variate
system is usually defined by the covariance matrix of the multi-variate
data. The linear transformation that diagonalizes the covariance matrix
can also be applied to the original data and produce a multi-variate
system with an inter-dimensional correlation of zero, i.e.; completely
uncorrelated multi-variate data. The linear transformation that
accomplishes this is the matrix of eigen-vectors or characteristic vectors.
A common application of this transformation is to reduce the
dimensionality of a multi-variate system. The objective is to summarize
most of the variance, or information content, in a system with a
lessor number of 'artificial' variates, i.e.; principal components.
Effectively, by uncorrelating the system, we are compressing most of the
information into a system with lower dimensionality.
Assume an n-variate system ( n channels of data ). Let K be the
n x n covariance matrix of this data and A be the matrix of eigen-vectors
of K. Associated with each eigen-vector A(j) there is an eigen-value or
characteristic root, e(j).
Let A(j) = (a(1j),a(2j),...,a(nj)) be the eigen-vector corresponding
to the jth largest eigen-value.
Let X = (x(1),x(2),...,x(n)) be the n-variate observation
( n-dimensional pixel ).
Then the jth principle component is: v(j) = A(j) X = a(ij)x(i)
NOTE: The jth eigen-value is actually the variance in the jth
principle componemt dimension. Therefore, the eigen-values are useful
as a measure the information content that can be expected in the output
pictures.
WRITTEN BY: J. D. Addington, 19 Oct. 1979
CURRENT COGNIZANT PROGRAMMER: Ron Alley
Made portable for UNIX W. Huey (CRI) 1994
Removed aliases and lets from tst pdf. FR76949 May 1995
PARAMETERS:
INP
input data sets
OUT
output data sets
SIZE
The standard Vicar size
field (sl,ss,nl,ns)
SL
Starting line
SS
Starting sample
NL
Number of lines
NS
Number of samples
LINC
Compute transform
from every nth line.
INC
Compute transform from every
nth line and nth sample
CORR
Compute eigenvectors from
the correlation matrix
DSTRETCH
Output decorrelation stretch
matrix, rather than eigenvector
matrix
DSCALE
(Used only with DSTRETCH)
Adjust the variance equalization
scaling factors by the specified
values.
EXCLUDE
Pixels with this DN in all
bands will be excluded from
all computations.
MSS
The number of input bands,
when the input is in MSS format.
USE
When input is MSS and not all
bands are to be used, use these
bands.
MSSO
The number of output components,
if the output is to be in MSS
format.
OUTPUT
The components to be output.
Default is to output the first
components, in order.
SAVE
The name for the parameter
dataset to hold the eigenvector
matrix.
TSAVE
The name for the parameter
dataset to hold the transpose
of the eigenvector matrix.
AREA
The subareas to be used to
compute statistics. Up to 50
regions (SL,SS,NL,NS) may be
entered. Default is to use
the entire image.
NVAR
FOR NON-IMAGE DATA ONLY
The number of dimensions
to the data.
DATA
FOR NON-IMAGE DATA ONLY
Up to 500 data values for
computing statistics.
See Examples:
Cognizant Programmer: