Skip to content

Click in-app to access the full platform documentation for your version of DataRobot.

Time series model package prediction intervals

Availability information

Time series prediction interval support for model packages, available as a public preview feature, is off by default. Contact your DataRobot representative or administrator for information on enabling the feature.

Feature flag: Enable computation of all Time-Series Intervals for .mlpkg

Now available for public preview, you can enable the computation of a model's time series prediction intervals (from 1 to 100) during model package generation. To run a DataRobot time series model in a remote prediction environment, you download a model package (.mlpkg file) from the model's deployment or the Leaderboard. In both locations, you can now choose to Compute prediction intervals during model package generation. You can then run prediction jobs with a portable prediction server (PPS) outside DataRobot.

Note

The Compute prediction intervals option is off by default because the computation and inclusion of prediction intervals can significantly increase the amount of time required to generate a model package.

Deployment model package download

To download a model package with prediction intervals from a deployment, ensure that your deployment supports model package downloads. The deployment must have a DataRobot build environment and an external prediction environment, which you can verify using the Governance Lens in the deployment inventory:

  1. In the external deployment, click Predictions > Portable Predictions.

  2. Click Compute prediction intervals, then click Download model package (.mlpkg).

    The download appears in the downloads bar when complete.

  3. Once the PPS download completes, use the provided code snippet to launch the Portable Prediction Server with the downloaded model package.

Leaderboard model package download

To download a model package with prediction intervals from a model in the Leaderboard, you can use the Predict > Deploy or Predict > Portable Predictions tab.

To download from the Predict > Deploy tab, take the following steps:

  1. Navigate to the model in the Leaderboard, then click Predict > Deploy.

  2. Click Compute prediction intervals, and then click Download .mlpkg.

    The download appears in the downloads bar when complete.

Availability information

The ability to download a model package from the Portable Predictions tab depends on the MLOps configuration for your organization.

To download from the Predict > Portable Predictions tab, take the following steps:

  1. Navigate to the model in the Leaderboard, then click Predict > Portable Predictions.

  2. Click Compute prediction intervals, and then click Download .mlpkg.

    The download appears in the downloads bar when complete.

  3. Once the PPS download completes, use the provided code snippet to launch the Portable Prediction Server with the downloaded model package.

PPS prediction interval configuration

After you've enabled prediction intervals for a model package and deployed the model to a portable prediction server, you can configure the prediction intervals percentile and exponential trend in the .yaml PPS configuration file or through the use of PPS environment variables. For more information on PPS configuration, see the Portable Prediction Server documentation.

Note

The environment variables below are only used if the YAML configuration isn't provided.

YAML Variable / Environment Variable Description Type Default
prediction_intervals_percentile / MLOPS_PREDICTION_INTERVALS_PERCENTILE Sets the percentile to use when defining the prediction interval range. integer 80
exponential_trend / MLOPS_EXPONENTIAL_TREND Controls whether DataRobot sets an exponential trend and applies a log transformation. boolean False

Updated June 23, 2022
Back to top