User-browser interaction-based fraud detection system
View Patent ↗Systems, methods and media for detecting fraudulent behavior during an Internet commerce session are disclosed. Embodiments of a method may include establishing an e-commerce session with a user and requesting fraud detection analysis of user-browser interaction during the e-commerce session. Embodiments may also include receiving fraud detection results for the e-commerce session, where the fraud detection results may provide an indication of a comparison between the user's interaction with a browser during the e-commerce session and known fraudulent behavior. Embodiments may also include performing an action based on the fraud detection results. Determining the fraud detection results may include determining user-browser interaction data associated with the session and comparing the user-browser interaction data to known fraudulent human or automated program behavior. Performing the action based on fraud detection results may include one or more of completing an e-commerce transaction, requesting additional authentication, denying the requested action, etc.
1. A method for detecting fraudulent behavior during an Internet commerce session, the method comprising:
establishing, by an application server computer system, an e-commerce session with a particular user;
requesting, by the application server computer system, fraud detection analysis of user-browser interaction during the e-commerce session;
determining, by the application server computer system, user-browser interaction data associated with the established e-commerce session with the particular user, the user-browser interaction data comprising indications of each of keyboard interactions, movement device interactions, and navigation/selection tendencies;
accessing, by a fraud detection computer system, stored user-browser interaction data for the particular user associated with previous e-commerce sessions by the particular user, the stored user-browser interaction data comprising indications of each of keyboard interactions, movement device interactions, and navigation/selection tendencies;
comparing, by the fraud detection computer system, the determined user-browser interaction data to both known fraudulent automated program behavior and to the accessed stored user-browser interaction data from previous e-commerce sessions by the particular user;
determining, by the fraud detection computer system, fraud detection results for the established e-commerce session based on the comparison of the determined user-browser interaction data with the known fraudulent automated program behavior and the stored user-browser interaction data from previous e-commerce sessions by the particular user;
transmitting, by the fraud detection computer system, the fraud detection results to the application server computer system; and
performing, by the application server computer system an action based on the fraud detection results.
2. A non-transitory machine-accessible medium of a non-transitory storage device, said non-transitory machine-accessible medium containing instructions effective, when executing in a data processing system, to cause said data processing system to perform operations comprising:
establishing, by an application server computer system, an e-commerce session with a particular user;
requesting, by the application server computer system, fraud detection analysis of user-browser interaction during the e-commerce session;
determining, by the application server computer system, user-browser interaction data associated with the established e-commerce session with the particular user, the user-browser interaction data comprising indications of each of keyboard interactions, movement device interactions, and navigation/selection tendencies;
accessing, by a fraud detection computer system, stored user-browser interaction data for the particular user associated with previous e-commerce sessions by the particular user, the stored user-browser interaction data comprising indications of each of keyboard interactions, movement device interactions, and navigation/selection tendencies;
comparing, by the fraud detection computer system, the determined user-browser interaction data to both known fraudulent automated program behavior and to the accessed stored user-browser interaction data from previous e-commerce sessions by the particular user;
determining, by the fraud detection computer system, fraud detection results for the established e-commerce session based on the comparison of the determined user-browser interaction data with the known fraudulent automated program behavior and the stored user-browser interaction data from previous e-commerce sessions by the particular user;
transmitting, by the fraud detection computer system, the fraud detection results to the application server computer system; and
performing, by the application server computer system an action based on the fraud detection results.
3. An e-commerce fraud detection system, the system comprising:
an application server computer system having one or more processors and in communication with an incoming interaction server and a fraud detection server computer system, the application server comprising:
an e-commerce application to establish a session with a particular user of a client computer system;
a fraud detection requester in communication with the e-commerce application to request fraud detection analysis of user-browser interaction for the established session and to receive fraud detection results for the established session; and
wherein the e-commerce application is adapted to perform an action based on the fraud detection results for the established session; and
a fraud detection server computer system having one or more processors and in communication with the application server computer system, the fraud detection server comprising:
an application server interface to receive from the application server computer system a request for fraud detection analysis of user-browser interaction for the established session with the particular user and to transmit fraud detection results to the application server computer system;
an interaction database interface to access stored user-browser interaction data associated with previous e-commerce sessions by the particular user, the stored user-browser interaction data comprising indications of each of keyboard interactions, movement device interactions, and navigation/selection tendencies;
a fraud detection module to determine user-browser interaction data associated with the established e-commerce session with the particular user, the user-browser interaction data comprising indications of each of keyboard interactions, movement device interactions, and navigation/selection tendencies; and
wherein the fraud detection module determines fraud detection results for the established e-commerce session based on a comparison of determined user-browser interaction data with both known fraudulent automated program behavior and the accessed stored user-browser interaction data from previous e-commerce sessions by the particular user.