How does the PUT endpoint work?

Basically, first, you run the normal Fraud API and the same transaction id that you use there, can be used on the PUT HTTP endpoint for the Label API and you can then say "Rejected" or "Risky" etc., this would provide our system with feedback, thus 'teaching' the Machine Learning algorithm to create better rules for you. 

Labeling can also be configured in the Settings part of the Admin panel to assign specific States (APPROVE or DECLINE) with certain Negative or Positive labels. 

You can send one or multiple transaction_ids in one PUT request to Label one or more transactions. You can read more about how the Label API works here.