I have a dataset which has ~60 columns and ~75000 rows. I used the isolation forest model to detect some outliers and do the shap plot to see the top features of these isolation forest.
My question is: Is it possible to use shap values as threshold for removing values? Like if I set the level of shap values for some certain variables, if shap values > or < threshold, they will be determined as outliers and can be removed. Is this executable? If so, how do I do this in python?