SYNOPSIS:
NormStream [options] Net [Input]
DESCRIPTION
The NormStream command is used to set coefficients of the linear
transformation between external data and input/output units activation
(see the Streams section). The 'Input' argument is optional only with
the -0 and -c options.
OPTIONS
SEE
ALSO
-S
This option changes the interpretation of the argument "Input". With the -S
option specified, the training data is taken from external files with the
basenames taken from the rows of the script file "Input". Otherwize "Input"
is itself a basename. In any case, the streams of the network read/write data
from files specified by the basename together with the information about file
extension and directory in each stream (see the Streams section).
-s stream
Process 'stream'. By default, all output streams are normalized, but this
option aloows you to specify particular streams. There can be any number of
-s options on the command line.
-d mult
By defalt, NormStream scans the training-data for the highest and the lowest
value for each component of the stream. Then it sets up the linear
transformation so that these values are mapped -1 and +1. This
normalization is obviously sensitive to the extreme values and therefore
it is sometimes better to use this option that normalizes based on the
mean and standard deviation. "mean - mult * stddev" is mapped to -1 and
"mean + mult * stddev" is mapped to +1.
-0
Reset the linear transformation of selected streams to the identity mapping.
-c a b
Set the coeficients of the linear transformation to 'a' and 'b' for all
components of all selected streams.