Dataset Actions and the Credit Score Target in DAI
- Predictive AI
- H2O Driverless AI
This lesson uses a loan application dataset to explain two foundational ideas in Driverless AI: the unit of analysis, here a single loan application, and the target column, here Credit_Score, which the platform trains models to predict. It describes Driverless AI's approach as running many candidate models in parallel, rather than testing one model at a time as you might in a notebook. It then walks through the action buttons available for any dataset, including Details for a data summary, Visualize for automatic charts, Data prep for transforming the data, and Predict, where an experiment gets set up and launched.
Key takeaways
- How the unit of analysis, a single loan application here, defines what each row in a dataset represents
- Why the target column, like Credit_Score, is the value Driverless AI trains models to predict
- What the Details, Visualize, and Data prep actions each do for a dataset before running an experiment
- How Driverless AI runs many candidate models in parallel instead of testing them one at a time
Start owning your AI