Smart cluster labeling using Generative AI¶
Access this AI accelerator on GitHub
The toughest part of unsupervised learning modeling is explaining clusters to end users. DataRobot not only allows users to build clustering models, but also provides insights per cluster to help users analyze the clusters. In most scenarios, the users building the models might not have the subject matter expertise to tailor the cluster labels towards the users consuming the models. This is where you can use Generative AI models to automatically label the clusters with some prompt engineering. Because the Generative AI models have been trained on vast amounts of domain and business datasets, they can understand and label the clusters tuned for end user expertise.
This AI Accelerator shows how to extract cluster insights from DataRobot models, use prompt engineering to label clusters, and then rename the clusters in the DataRobot project.
You will explore the following:
- Use the API to extract Cluster Insights from DataRobot unsupervised learning projects.
- Use Generative AI and Prompt Engineering to consume cluster insights and create cluster labels for DataRobot clusters.
- Use the API to rename DataRobot clusters.