Score the percentage of money lost of the total transaction amount process due to fraud.
Create instance of FraudCost
Calculate the percent difference between scores.
Determine if a transaction is fraud given predicted probabilities, threshold, and dataframe with transaction amount.
Calculate amount lost to fraud per transaction given predictions, true values, and dataframe with transaction amount.
Learn a binary classification threshold which optimizes the current objective.
Returns a numerical score indicating performance based on the differences between the predicted and actual values.
Validates the input based on a few simple checks.