Setting Up a Production Model Scoring Process

1. Model Approval & Production

First step is to approve and implement a production model version for the predictive model.  See the article for how to approve and implement the model version you want to use for scoring.


2. Scoring Catalog

The second step is to ensure you have setup an appropriate scoring catalog to hold the scores.  The catalog must be set to use primary keys that match the dataset you will be scoring.  If you do not already have a scoring catalog setup, see to set one up before continuing.


3. Create Scoring Job

Now you have to create and setup a scoring job. This is where you will connect a model to a dataset, and ultimately to a scoring catalog where scores will be stored.

See for creating a scoring job and for how to edit the settings.


4. Manually Run or (optionally) Automate the Scoring Job

Once setup, a scoring job can be run at any time manually (see, or set to run automatically (see  Automating the scoring job will ensure it always runs when the underlying dataset is refreshed.


5. View and (optionally) Export Scores

When the scoring job finishes executing, the scoring catalog will be populated with the results.  You will likely want to at least view the scores or possibly export them for use in other applications.