Model validation flow

The following steps summarize the flow of validating a model in H2O Model Validation:

• Step 1: Establish a Connection and import a model
• Step 2: Create validation test(s)
• Step 3: Review test(s) metrics

For more information on each of the preceding steps, refer to the following sections.

Step 1: Establish a Connection and import a model

Establish a Connection that lets you import your models from one of the supported platforms. Once a Connection has been established, you can begin importing models.

• To learn how to establish a Connection, see Create a Connection.
• To learn how to import a model from an established Connection, see Import a model.
• To learn about supported platforms, see Supported platforms.

Step 2: Create validation test(s)

Run one or multiple validation tests for the imported model. H2O Model Validation offers a variety of validation tests to analyze the robustness and stability of your machine learning models. You can fine-tune the parameters of each test by using the available configuration settings.

• Validation tests for a model:
  • Adversarial similarity
    • To learn how to create an adversarial similarity test, see Create an adversarial similarity test.
  • Backtesting
    • To learn how to create a backtesting test, see Create a backtesting test.
  • Drift setection
    • To learn how to create a drift detection test, see Create a drift detection test.
  • Size sependency
    • To learn how to create a size dependency test, see Create a size dependency test.

Step 3: Review test(s) metrics

Review the validation test(s) metrics. For each validation test, H2O Model Validation offers a variety of metrics in graphs, charts, and heatmaps that let you assess the results of each test in detail.

• Metrics: Adversarial similarity
• Metrics: Backtesting
• Metrics: Drift detection
• Metrics: Size dependency

---

Feedback

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