The random seed represents the starting point that
IBM ®
Campaign uses to select records
randomly.
About this task
Some processes provide the ability to select a random set of records. If you decide to
select random records, there are times when you might want to change the random seed. For
example:
- Your random sample is producing highly skewed results (for example, if all males in your
data fall into one group and all females into another).
- You have the same number of records in the same sequence, and using the same seed value each
time that you run this process results in records being created into the same samples.
Follow the steps below to generate a different starting point for random record
selection.
Procedure
-
Click the Cell Size Limit tab of the Audience, Extract, Merge,
Select, or Sample configuration dialog.
-
Perform one of the following operations:
- Click the Pick button next to the Random
Seed option to randomly select a new seed value.
The same random set of records
will be used for each subsequent run of the process (unless the input to the process
changes). This is important if you intend to use the process results for modeling purposes,
because different modeling algorithms must be compared across the same set of records to
determine each model's effectiveness.
- If you do not intend to use the results for modeling, you can make the process select a
different random set of records each time it runs. To do this, enter a zero (0) as the
Random Seed. A value of 0 ensures that a different random set of
records will be selected each time the process runs.
Tip: The Segment process does not include a Random Seed control. To make the
Segment process select a different random set of records each time it runs, set the Random
Seed to 0 in the upstream process that provides input to the Segment process.