The BehavioCheckout demonstrates a sample shop, and you will need to register a user to be able to access the application. It is intended to show the capabilities to authenticate user's by their unique mobile keyboard behavior.
The software gathers your behaviometric profile when you type in the given fields. It then, on submit, sends the data to a BehavioWeb server for evaluation.
The behavioral features we are taking into account are: keystroke dynamics, pressure, touches, accelerometer and gyroscope
You will first need to train the behavior of the user, this is done by signing in with your registered username/password, then filling the gift page as well as the checkout page, you will then be presented with a score page. The score page will tell you the results of the transactions. Training means that you need to input more transactions before you are evaluated. Green means the behavior was correct to the user profile, Yellow means the behavior is uncertain, Red means the behavior was incorrect.
From the 11th transaction on, your behavior will be scored against your trained profile. So try it a couple of times and observe the results. You should see a green result page. Hand it then over to another user and let him try to login with your username password combo.
Keep in mind, that this is not the usual way like users encountering the system. In a production environment, the system is learning the user behavior always in the background and is trained already before a transaction is been scored.
The score you will see is the similarity of the actual user, to the stored behavior from the profile. The closer the score is to 100%, the more probable it is that the sample comes from the correct user. The confidence is showing how confident the system is in his decision. This depends on the amount of data in the profile, the stability of the user behavior as well as on the amount of data in the current transaction.