Time-aware experiments¶
Experiments are the individual "projects" within a Use Case. They allow you to vary data, targets, and modeling settings to find the optimal models to solve your business problem. Within each experiment, you have access to its Leaderboard and model insights, as well as information that summarizes the data and experiment setup.
There are two types of AI experiments available in Workbench:
-
Time-aware modeling, described on these pages, uses date/time partitioning to model using time-relevant data to make row-by-row predictions, time series forecasts, or current value predictions ("nowcasts").
-
Predictive modeling, makes row-by-row predictions based on your data.
The following sections help to understand building time-aware experiments in Workbench:
| Topic | Description |
|---|---|
| Create experiments | |
| Supervised experiment setup | Specify a target to build models that predict or forecast time-relevant data. Supervised learning uses the other features in your dataset to make forecasts or predictions. |
| Unsupervised experiment setup | Build clustering models that surface insights about patterns in your data or anomaly detection to identify outliers. Unsupervised learning uses unlabeled data to surface insights about patterns in your data, answering questions like "Are there anomalies in my data?" |
| Manage models | |
| Manage the Leaderboard | Navigate and filter the Leaderboard; create feature lists. |
| Add/retrain models | Retrain existing models and add models from the blueprint repository. |
| Explore model insights | |
| Evaluate models | View model insights to help evaluate models. |
| Reference | |
| Time series reference | Read reference material explaining the DataRobot framework for implementing time series modeling. |
Feature considerations¶
Consider the following when working with date/time partitioned projects in Workbench:
- You cannot create feature lists on derived features.
- Model comparison is not supported.
- Eureqa model insights are not available.