Choosing a Problem Type in H2O Driverless AI
- Predictive AI
- H2O Driverless AI
This lesson introduces the machine learning problem types supported by H2O Driverless AI, split into supervised methods such as classification, regression, time series, and NLP, and unsupervised methods such as clustering. It focuses on a supervised learning example, using historical loan application data to predict a credit score, and explains the difference between a regression problem, predicting a number, and a classification problem, predicting a category. The lesson notes that Driverless AI also supports more advanced time series, image, and NLP use cases, which are covered in separate courses on the H2O University portal.
Key takeaways
- How a supervised learning problem differs from an unsupervised one, based on whether the target has a label
- What separates a regression problem, predicting a number, from a classification problem, predicting a category
- The credit score example used in this course, predicting whether a loan application is good or bad based on historical data
- Why more advanced problem types like time series, image recognition, and NLP sit outside this introductory course
Start owning your AI