Activate blackbox scoring

Once activated, blackbox machine learning fraud probabilities are layered into the decision powered by the Scoring Engine. You will see the probability (Blackbox score) returned by the algorithm, alongside a fraud score calculated based on your existing Default and Custom rules.

If you are just about to start using SEON, you will find blackbox machine learning enabled by default in your new account. You don't have to do anything but start using it and setting up Feedback Loops & Labels to ensure your machine learning model is trained on both bad and good examples.

If you are just about to start using SEON, you will find blackbox machine learning enabled by default in your new account. The blackbox machine learning algorithm calculates the fraud probability for every transaction. A customer-specific prediction model becomes automatically available to you if you have over 1000 transactions, of which at least 100 have the DECLINE state. You can turn the blackbox scoring feature off anytime. If you can't find blackbox score settings under your Settings tab, reach out to our Customer Success team.

1. Go to Settings.
2. Head to the Machine Learning tab.
3. Scroll down to the Blackbox Score section.
4. Enable with the toggle switch left of Calculate Blackbox score for every transaction.

If you want to, you can also set up a custom threshold for approving or declining transactions automatically.

You will also receive this score in the Fraud API response by default.