inputWhiten
Whitening transformation based on eigen decomposition
Contents
Syntax
- X2=inputWhiten(X)
- X2=inputWhiten(X, plotOpt)
- [X2, meanVec, A]=inputWhiten(...)
Description
[X2, meanVec, A]=inputWhiten(X) takes the input data matrix X (with each column being an observation vector) and returns the whitened data matrix X2 which has a identity covariance matrix, with meanVec being the mean vector of the original data matrix X, and A being the transformation matrix.
For any given data matrix Y to go through the same whitening process, the code is like this:
Y2=A*(Y-meanVec*ones(1, size(Y,2)))
References
- [1] R.O. Duda, P.E. Hart, and D.G. Stork, Pattern Classification,
- New York: John Wiley & Sons, 2001, pp. 34.
Example
dataNum=1000; X=randn(2,dataNum); meanVec=[20; -10]; A=[9, 6; 3, 1]; X=A*X+meanVec*ones(1,dataNum); X2=inputWhiten(X, 1);
