Compare two models
TL;DR: You can use Evidently to evaluate the performance of two models on the same data (in test / shadow deployment). Here is an example Jupyter notebook.

Tutorial - "What Is Your Model Hiding?"

In this tutorial, we train two classification models that predict employee attrition. Then, we compare and evaluate the differences in their performance despite the similar ROC AUC. The tutorial uses the Probabilistic Classification report.
Last modified 6d ago
Copy link
Edit on GitHub