Robustness
Overview
H2O Model Validation's robustness test is designed to assess the robustness of a machine learning model against noisy datasets. The test perturbs the dataset with various methods to introduce noise based on the type of each column in the dataset.
To conduct the test, H2O Model Validation uses a combination of customizable settings to customize the perturbation process.
By conducting a robustness test with H2O Model Validation, you can gain a better understanding of how well your machine learning model will perform in real-world scenarios where the data may be noisy or corrupted. This test can help you identify potential issues and take steps to address them before deploying the model in production (for example, clean the data, normalize the data, apply feature scaling, or use a different model that is more robust to noise).
Resources
- To learn how to create a robustness test, see Create a robustness test.
- See Settings: Robustness to learn about all the settings for a robustness test.
- See Metrics: Robustness to learn about all the metrics for a robustness test.
- Submit and view feedback for this page
- Send feedback about H2O Model Validation to cloud-feedback@h2o.ai