In compare and create results of classification.

Currently, PyCaret shows a weighted average for model performance, but this doesn't always tell the full story. For example, if you have two classes like "yes" and "no," and the model performs really well for "no" but poorly for "yes," the weighted score might still look good. This can hide the fact that the model isn't doing well on the minority class ("yes").

The problem is, you might end up selecting and deploying the model that looks the best based on these weighted scores, but in reality, it’s the worst model for catching the minority class. What you really need is to focus on metrics that show the performance of each class, especially the minority ones, so you can truly understand how well the model is working for all your data.

Describe your proposed solution

Best will be to show results of all classes which is like shown in classification report of sklearn.

Describe alternatives you've considered, if relevant

No response

Additional context

No response