Scenarios
A scenario is a prediction task built on top of your foundation model. You define what to predict with a target function, choose how to predict it with a task type, then train, evaluate, and run inference.
Prerequisite
You need a trained foundation model before building scenarios. See Foundation Model.
In this section
| Guide | Description |
|---|---|
| Tasks & Targets | Select a task type and write a matching target function |
| Model Configuration | Configure task-specific model settings and run training |
| Retraining on New Data | Refresh or continue training a scenario on newer data by passing new date ranges |
| Testing Against Ground Truth | Test the trained model and review metrics |
| Inference | Generate predictions on new data |
| Interpretation | Inspect prediction drivers at the entity or global level |