Arize AI
Search…
6. Troubleshoot Drift
Drift Tab offers a deep dive into Prediction Drift, Feature Drift, and Data Distribution Change.
Your model's prediction drift monitor goes off! How do you troubleshoot what is causing your outputs to drift?

Navigate To The Drift Tab

From the 'Space Overview' page, click on your model of interest. From the 'Model Overview' page, click the 'Drift' tab located on the top navigation bar.
Navigate to the 'Drift Tab' from the 'Space Overview' page
If you're looking at individual monitors in the 'Monitors' tab, access the 'Drift' tab from the 'Troubleshoot Drift' button located on the top right of your drift monitor.
Navigate to the Drift Page from a monitor

Prediction Drift

Prediction Drift refers to a change in the prediction distribution generated by your model. Arize can surface a change in distributions for all model types and problems such as regression, binary, multi-classification, and score models.
Distribution Comparisons for values above and below the model baseline
By clicking on specific points of the Prediction Drift curve, Arize also surfaces changes in Distribution Comparison and Feature Drift. This allows you to narrow down on issues to troubleshoot your model.

Feature Drift

Feature Drift refers to changes in your feature distribution. Feature drift is relatively common in the real world and can happen at any time in production while remaining undetected.
Arize allows you to analyze any situations pertinent to your model. You can sort in any of the columns to deep drive into explaining your model performance.
Prediction Drift Impact can surface when drift has impacted your model. Drift (PSI) is a measurement of how much your distribution has drifted. Lastly, Feature Importance helps you explain why even a small Drift (PSI) can have a significant Drift Impact.

Troubleshooting Tutorial

For a quick-start intro on how to use the drift tab, check out the Troubleshooting Drift Tutorial!
Google Colaboratory

Additional Resources

Questions? Email us at [email protected] or Slack us in the #arize-support channel