Model Comparison Results Windows

Assessment

The assessment plots available depend on the models being compared. For classification models, the plots displayed are Lift, ROC, and Misclassification. For numerical models, the plots displayed are observed response value and predicted response value.

Fit Statistic

The Fit Statistic plot displays the criterion specified in the Fit statistic property. In the following image, the observed average value is plotted for a linear regression and a GLM model. The champion model is indicated in the plot. It is displayed different from the other models.
Fit Statistic Plot

Details Table

When you click the Details View Icon icon at the top of the model pane, the details panel is displayed at the bottom of the model pane. The details table contains the following information:
Statistics
Provides summary statistics for each model in the comparison. The value in the Selected column, either Yes or No, indicates which model the model comparison tool prefers based on the criterion specified in the Fit statistic property. However, the statistics listed in the details table can differ from those listed in the Fit statistic property.
Variable Importance
Indicates which variables had the greatest impact on each of the models in the comparison.