How to tune hyperparameters in TIBCO Data Science Team Studio?
Article ID: KB0073834
Updated On:
| Products | Versions |
|---|---|
| Spotfire Data Science Team Studio | 6.5.0 |
Description
How do we tune hyperparameters in TIBCO Data Science Team Studio?
Issue/Introduction
We can tune hyperparameters in Team Studio either with a simple workflow or a python code.
Environment
Operating System: Linux
Resolution
Approach 1: Build a Workflow
We use cross-validation sets and grid search to find the best parameter in some of the AutoML operators. They are tuned in the modeling operators: AutoML logistic regression, AutoML Random Forest, and AutoML Gradient boosting trees. We can achieve the same with other regular operators by using the same ML operator multiple times in a workflow and passing a different range of values to get the best possible output.
Approach 2: Python code
We can write a python code in Team Studio's python notebook that can find hyperparameters and call them in the workflow using a Python Execute operator.
Additional Information
We have attached a sample workflow and notebook for your reference.
Attachments
Feedback
Yes
No
Powered by
