what's the difference of Replacement and Impute node in em7.1?

[TomiKong]

When we should use Replacement node and when use Impute node?

Thanks.

Best

Tom

[jwexler · Accepted solution]

Tom, at a high-level the Replacement node allows you to assign/replace or trim values in your data.  For example you may want to trim the tails of your distribution to tighten your analysis.  You could then use an Impute node for example.  The Impute node essentially replaces missing values in your data set.  As you may know, if values are missing in your modeling set, observations are often excluded from the analysis.

Documentation can be found here:

http://support.sas.com/documentation/solutions/miner/em/emref.pdf

Thanks,

Jonathan