From Turbo Prep
Handle
Numbers
Perform
Normalization
Information:
Normalization is a common technique which ensures that all numeric columns of
your data set are roughly one the same scale. Each column is rescaled so that
the average of the resulting column is 0 and the standard deviation for all
columns becomes 1. By doing this, different scales won't impact machine
learning models which is in particular important for distance-based methods.
However, the resulting models are somewhat harder to interpret since the scales
have changes to something which does not occur in reality. If you use Auto Model, it is
usually better to let Auto Model do the normalizations only when they are
necessary.
What step am I missing where that happens in Auto Model?
