Skip to content

On-premise users: click in-app to access the full platform documentation for your version of DataRobot.

NextGen experience > AI experimentation > Time-aware experiments > Create time-aware experiments

Create time-aware experiments

Time-aware modeling is based on Date/time as the partitioning method. Supervised learning uses the other features in your dataset to make forecasts and predictions. Unsupervised learning, by contrast, uses unlabeled data to surface insights about patterns in your data.

The following types of time-aware modeling are available. For all supervised learning methods, use the basic setup to create an experiment.

Type Description Usage
Time-aware predictions (Supervised) Assigns rows to backtests chronologically and makes row-by-row predictions. Provides no feature engineering. Use when forecasting is not needed.
Time-aware predictions with feature transformations (Supervised) Assigns rows by forecast distance, builds separate models for each distance, and then makes row-by-row predictions. Recommended to combine with time series wrangling, which provides transparent and flexible feature engineering. Use when:
  • Forecasting is not needed, but you want predictions based on forecast distance.
  • Dataset is larger than 10GB.
  • Full transparency of the transformation process is desired.
.
Time series forecasts (Supervised) Forecasts multiple future values of the target using the DataRobot automated feature derivation process to create the time series modeling dataset. Use when:
  • Dataset is smaller than 10GB.
  • Full transformation process, preprocessor -> extractor -> postprocessor, for all features is required.
Unsupervised experiment setup Builds clustering models that surface insights about patterns in your data or performs anomaly detection to identify outliers. Use without specifying a target.

Note

There is extensive material available about the fundamentals of time-aware modeling. While those instructions largely represent the workflow as applied in DataRobot Classic, the reference material describing the framework, feature derivation process for time series forecasts, and more are still applicable.

Updated January 21, 2025