Compare models

Overview

H2O Model Validation enables you to compare models to understand the similarities and differences between models.

note

H2O Model Validation utilizes the model's summaries to compare the models.

Instructions

To compare models, consider the following instructions:

1. In the H2O Model Validation navigation menu, click Models.
2. Click the Select models toggle.
3. In the models table, select at least two models to compare.
4. Click Compare.
   note

   • In the Model summary tab, a comparison table and a feature importance chart appear.
     • To learn more, see:
       • Comparison table
       • Chart: Feature importance
   • In the other tabs, you can observe the comparison metrics of certain tests associated with the selected models.
     • To learn more, see Comparison metrics: Validation tests

Model summary tab

Comparison table

Column name	Description
Model summary name	The name of the model's summary.
Train dataset name	Name of the experiment train dataset.
Validation dataset name	Name of the experiment validation dataset.
Test dataset name	Name of the experiment test dataset.
Scorer	The scorer of the experiment.
Validation score	Experiment validation score value.
Test score	Experiment test score value.
Accuracy	Experiment accuracy value.
Time	Experiment time value.
Interpretability	Experiment interpretability value.
Task	Experiment problem type (for example, regression).
Target	Experiment target column (target feature).
Dropped columns	Dropped columns that Driverless AI dropped during the experiment to not use as predictors.

Chart: Feature importance

The Feature importance chart displays all the features of the compared models.

• X-axis: Feature names
• Y-axis: Gain values (the importance of the feature in the model)

---

Feedback

• Submit and view feedback for this page
• Send feedback about H2O Model Validation to cloud-feedback@h2o.ai