whitening | Whitening transformation |
-d <data> | data frame to be preprocessed |
-dout <dataout> | name of the preprocessed data frame |
The whitening transformation is a linear preprocessing method which allows correlation among variables. The transformation is performed using eigenvectors from covariance matrix in a way that is illustrated roughly in the figure below. Data frame to be processed may only contain floating point vectors. This transformation is useful, if source data contains fields, which differ a lot in magnitude.
Example (ex4.12): Load some arbitrary data and perform the whitening transformation.
# Load some data in NDA> load sin.dat # Perform the whitening transformation NDA> whitening -d sin -dout whitesin