Hello there!

If you're like me, and work in the finance / banking industry, then you'll understand the rigor of going through model validation. Feature selection is one of the key sections in every model documentation. Hopefully, this code will provide a generalizable way to identify which features to keep.

Use the RecursiveFeatureSelection class to iteratively remove less important features with each iteration. Then, use the Visualizers to show the model performance as more and more features are removed.

Now, you may ask: "Why don't you just use sklearn.feature_selection.RFECV?"

• RFECV provides less transparency as it only gives you one model in the end with the optimal number of features. However, you may want more freedom to choose for yourself.
• Most model validation departments will want you to "show" your work.

Currently, this only supports binary classification, as that is the bulk of my models.