Today, we’d like to look at how users of both H2O.ai and explainX can benefit from a new integration that enables H2O AutoML library to be used in explainX open-source platform. H2O.ai users can leverage explainX in a workflow to provide model explainability (currently only available in H2O Driverless AI, the enterprise flagship product of H2O), model debugging and model explanatory analysis.
As this is the first release, the features may be limited but over the course of next few months, we will consistently add further support for H2O users leveraging explainX for model explainability. The aim of this article is to provide equip you with the six steps for getting started. We have already added the code for you and all you need to do is follow along!
The 6 steps to get started with the explainX and H2O are:
1. Import required dependencies
2. Run H2O server
3. Transform data in H2O Data Frame
4. Train H2O model
5. Call the explainX function by passing in the H2O model and your testing dataset
6. Access the interactive explainX dashboard within your workspace!
And voila! You're good to go.
Here is a video tutorial that covers the entire process:
We'd recommend you dive deeper into our documentation to learn about each techniques used and how you can leverage it to understand and explain the model behavior.
The expanded integration between H2O.ai and explainX brings together all-encompassing, intuitive, automated machine learning from H2O.ai with the model explainability from explainX. We're very excited to further enhance these capabilities and also support ensembles models within the H2O AutoML library.
Head over to our Github and start using the explainX library today!