PLS_Toolbox Documentation: snv	< simpls	spcreadr >

snv

Purpose

Standard Normal Variate scaling.

Synopsis

[xcorr,mns,sds] = snv(x);      %perform snv scaling

x = snv(xcorr,mns,sds);        %undo snv

Description

Scales rows of the input x to be mean zero and unit standard deviation. This is the same as autoscaling the transpose of x.

INPUT:

                         x =   M by N matrix of data to be scaled (class "double" or "dataset").

OPTIONAL INPUTS:

                     mns =   a vector of length M of means, and

                     sds =   vector of length M of standard deviations.

OUTPUTS:

                 xcorr =   the scaled data (xcorr will be the same class as x),

                     mns =   vector of means for each row, and

          sds =   vector of standard deviations for each row.

To rescale or "undo" SNV, inputs are xcorr, mns, and sds from a previous SNV call. The output will be the original x.

See Also

auto, normaliz, preprocess