Test and validate machine learning models

In the testing phase, you test the model that you trained to see whether the model produces the outcome that you want and to check the accuracy of the model. If the model meets your criteria, you can deploy it to production. If it doesn't pass the tests, it needs better data, and then should be retested until you get the desired output. Model testing allows you to evaluate the quality of your data, algorithms, or parameters before you deploy the model.

There can be several steps when testing your model, including:

1. Running the notebook to set up the monitors
2. Evaluating the model
3. Observing the model monitors for quality
4. Observing the model monitors for fairness
5. Observing the model monitors for explainability
6. Promoting the model to preproduction and approving the model

Try it out

To learn more about testing models in Watson Studio and to try it out, take a look at the AI Governance example project.

The tutorials in this project let you build, deploy, and track a model with an AI governance use case of a data fabric trial. The goal of the example is to train a model to predict which applicants qualify for mortgages and then deploy the model for evaluation. You also test and validate the model for fairness, accuracy, trustworthiness, and readiness for production.

Summary

This section provided an overview the testing phase for machine learning model development. For more information on testing with the various machine learning model creation options in Watson Studio, see Getting started with building, deploying, and trusting models.