| CPC G06Q 20/409 (2013.01) | 20 Claims |

|
1. A computer-implemented method for device fingerprinting, the method comprising:
matching device configuration data associated with a plurality of devices including a client device, wherein the device configuration data includes data for the client device, and wherein matching includes using a time period to generate a unique set of device configuration characteristics associated with the time period;
training a learning system including a neural network to create one or more datasets that result in uniqueness for devices;
identifying, via the neural network, one or more device configuration characteristics that create a non-unique overlap between device configuration characteristics associated with one or more other devices of the plurality of devices;
reducing memory use within a server computer and reducing opportunities for spoofing by removing the one or more device configuration characteristics from the set of device configuration characteristics, wherein removing includes associating the device configuration characteristics with a deletion timer and generating a unique set of device configuration characteristics that are deleted when the deletion timer expires without additional device configuration characteristics being received by the client device;
receiving, at the server computer, transaction request data associated with the client device;
verifying the client device using the unique set of device configuration characteristics; and
facilitating a transaction associated with the transaction request data based on verification of the client device using the unique set of device configuration characteristics.
|