Model Interpretation

This chapter describes Machine Learning Interpretability (MLI) in Driverless AI.

Notes:

• MLI no longer runs on experiments from previous releases, but also does not require an internet connection to run on current models.
• MLI is not supported for NLP experiments or for multiclass Time Series experiments. Contact H2O support for assistance on interpreting NLP models.

Additional Resources

• Click here to download our MLI cheat sheet.
• Click here to view our H2O Driverless AI Machine Learning Interpretability walkthrough video.

• The Interpreted Models Page
• MLI for Non-Time-Series Experiments
• MLI for Time-Series Experiments
• General Considerations