Driverless AI: Setting Up a Prediction Experiment
- Predictive AI
- H2O Driverless AI
This lesson walks through launching a first experiment in H2O Driverless AI using a credit score dataset to predict whether a score will be good or bad. It covers selecting the training and test datasets, naming the experiment, and the optional fields for a validation dataset, fold column, weight column, and time column that appear during setup. The lesson also points out the Assistant button, which adds an interactive tour with pop up explanations for each part of the experiment setup page for first time users.
Key takeaways
- How training and test datasets are each assigned separately before an experiment is launched
- What the optional fold, weight, and time column fields control during experiment setup
- Why selecting a target column with only two unique values sets up a supervised classification experiment
- How the Assistant button adds an interactive walkthrough of each section on the experiment setup page
Start owning your AI