Running training and scoring jobs
Complete the following steps to run the training and scoring jobs for data science models.
Run the training job
You can do the following to run the training job.
To run the training job:
-
Click the Oracle icon
in the bottom-right corner to open the navigation menu.
- Select Intelligence workbench.
- Click the Action menu (
) for the data science model you want to analyze and select Start Training. Monitor the status of the job from the Jobs dashboard.
When the job completes, Oracle Unity will automatically run the Scoring job at 8 a.m. UTC every day. This job will predict values for all new records ingested in the past 24 hours. If needed, you can manually run the scoring job.
Run the scoring job
You can do the following to run the scoring job.
To run the scoring job:
-
Click the Oracle icon in the bottom-right corner to open the navigation menu.
- Select Intelligence workbench.
- Click the Action menu () for the Scoring job you want to run and select Start Scoring. The job runs in three stages and takes about an hour to complete.
- Monitor the status of the job from the Jobs dashboard. When the job successfully completes, move on to the next step.
-
Click the Oracle icon in the bottom-right corner to open the navigation menu.
- Select Data feeds.
- Click the Action menu () for the System - Data warehouse job and select Start.
- Click the Action menu () for the System - Identity resolution pipeline job and select Start.
When all the jobs complete, you can start using and accessing the output data.