Skip to main content
Version: v0.17.0

Metrics: Segment performance

Overview​

H2O Model Validation offers an array of metrics to understand a segment performance test. Below, each metric is described in turn.

Settings​

{Metric} by segment (split by 2 features)​

The {metric} performance by segment (spli by 2 features) bubble graph displays the data segments of 1 or 2 defined variables. {Metric} refers to the model's scorer.

  • X-axis: A variable's value bins
  • Y-axis: A variable's value bins
  • Bubble size: Size of the data segment
  • Bubble color: The estimated model error using the metric selected in the test settings

Bubble graph: Performance by segment (split by 2 features)

Features to display​

In the features to display card, you can define the X and Y-axis for the following graph: {Metric} by segment (split by 2 features).

eatures to display

Features by importance for model performance segmentation​

The feature by importance for model performance segmentation chart displays the rankings of the model variables in terms of their value for the segment performance analysis.

Table: Features by importance for model performance segmentation

Feedback