日本語
DataRobot is designed to help you experiment with different modeling approaches, data preparation techniques, and problem framings. You can iterate fast with a tight feedback loop to quickly arrive at the best approach.
Sometimes you may wish to break your use case into multiple models, likely across multiple DataRobot projects. Maybe you want to build a separate model for each country or one for different periods of the year. In this case, it helps to bring all of your model performances and insights into one chart.
About this Accelerator
This accelerator shares several Python functions which can take the DataRobot insights - specifically model error, feature effects (partial dependence), and feature importance (Shap or permutation-based) and bring them together into one chart, allowing you to understand all of your models in one place and more easily share your findings with stakeholders
What you will learn
- Setup: import libraries and connect to DataRobot
- Accuracy Python function
- Feature Impact Python function
- Feature Effects Python function
- Example use and outputs
