Filter Your Training and Scoring Datasets

Focus your predictions on a specific subset of the object data by defining filter conditions. Get more useful predictions by excluding records irrelevant to your prediction. Ensure that you have sufficient records in your training and scoring datasets by counting the records. Filtering is optional. You can use the complete dataset for training and scoring.

Required Editions

Available in: Lightning Experience

User Permissions Needed
To set up a CRM Analytics template configuration:	Scoring Framework Admin

1. From Setup, in the Quick Find box, enter Industries Cloud Einstein, and then select Scoring Framework.
2. On the card of the template configuration that you want to use, click , and select Edit.
3. For Filter Your Training and Scoring Datasets, click Set Up.
4. Add conditions using the object’s fields to filter your training and scoring datasets.
   The same set of conditions is used to filter both the training and scoring datasets.
5. To determine the number of training and scoring records, click Count Records.
   If you have an insufficient number of records, edit your conditions, and recount the number of example and scoring records.
6. To continue defining the template configuration, click Save & Continue.
   You can also save your changes and return to the Scoring Framework Setup page. If you don’t want to filter training and scoring datasets, click Clear Selections.

Example

A financial institution uses Scoring Framework to get predictions about customers in the US who are likely to purchase more assets. The financial institution’s Salesforce admin adds conditions to filter accounts with the billing country as US.