Hints

How should I use a CounterFactual result when updating my data features for training my AI model?

How can I use the result of SHAP values to choose input features to my AI model?

Try removing some features that SHAP says is dragging down the price, but be careful with removing the categories as they also seem to help the prediction in some cases.

When you change the features, the datapoints that predict below zero might also change, so make sure to change the DATAPOINT that’s used for CounterFactual and SHAP.

Remove the discounted and discount_amount features to get only positive values.