Batch prediction methods¶
DataRobot offers a variety of methods to efficiently score large files via batch predictions:
| Method | Description |
|---|---|
| Batch Prediction UI | Configure Batch Prediction jobs directly from the DataRobot interface included in deployments via MLOps. |
| Batch Prediction API | Use the API to create Batch Prediction jobs that score to and from local files, S3 buckets, the AI Catalog, and databases. |
| Batch prediction scripts | Use command-line tools that wrap the Batch Prediction API, available for Windows, macOS, and Linux. |
In addition, you can monitor and manage predictions using monitoring jobs and batch prediction jobs:
| Method | Description |
|---|---|
| Monitoring Jobs UI | Use the job definition UI to create monitoring jobs, allowing DataRobot to monitor deployments running and storing feature data and predictions outside of DataRobot. |
| Monitoring Jobs API | Use the Batch Monitoring API to create monitoring job definitions, allowing DataRobot to monitor deployments running and storing feature data and predictions outside of DataRobot. |
| Batch Jobs tab | Use the Batch Jobs tab to view and manage monitoring and prediction jobs. |