The following sections describe the components to making predictions in DataRobot:
- Generate predictions: Initiate a prediction job with the
Model.request_predictions()method. This method can use either a training dataset or predictions dataset for scoring.
- Batch predictions: Score large sets of data with batch predictions. You can define jobs and their schedule.
- Prediction API: Use DataRobot’s Prediction API. to make predictions on both a dedicated and/or a standalone prediction server.
- Scoring Code: Qualifying models allow you to export Scoring Code and use DataRobot-generated models outside of the platform
- Batch Predictions
- Make batch predictions with a deployment
- Scoring local CSV files
- Scoring from and to S3
- Scoring from and to Azure Cloud Storage
- Scoring from and to Google Cloud Platform
- Wiring a Batch Prediction Job manually
- Copying a previously submitted job
- Scoring an in-memory Pandas DataFrame
- Batch Prediction Job Definitions
- Job Definitions
- Executing a job definition