The most comprehensive bank Interview Answer Review tooling available online.

False positives: This often referred to known as Type 1 error .This indicates that my formulated null hypothesis (H0) is true but I reject it( My decision about estimation of population parameters are correct but I can't accept (reject) .In contrast Type 2 error also known as false negatives this represent my null hypothesis is false but I failed to reject it ,(I don't have any strong enough evidence to the reject this ).

I think any decision tree approach would be best to handle high cardinal data. We can use random forest for a relatively interpretable approach otherwise we can use XGBoost for a more robust and accurate approach.