US 12,386,721 B2
Anomaly detection using user behavioral biometrics profiling method and apparatus
Bharatwaaj Shankar, Chennai (IN); Ishi Khamesra, Rajasthan (IN); Subham Biswas, Maharashtra (IN); Anne Lourdu Grace Joseph Martin, Chennai (IN); Malaya Rout, Tamil Nadu (IN); Venkadesh X. Thirupathi, Chennai (IN); Navin Dalavai Premkumar, Tamil Nadu (IN); and Mohit Kumar, Kalyan Nagar Post (IN)
Assigned to Verizon Patent and Licensing Inc., Basking Ridge, NJ (US)
Filed by VERIZON PATENT AND LICENSING INC., Basking Ridge, NJ (US)
Filed on Aug. 4, 2021, as Appl. No. 17/393,517.
Prior Publication US 2023/0043793 A1, Feb. 9, 2023
Int. Cl. G06F 11/34 (2006.01); G06N 20/10 (2019.01); H04L 67/50 (2022.01)
CPC G06F 11/3438 (2013.01) [G06F 11/3466 (2013.01); G06N 20/10 (2019.01); H04L 67/535 (2022.05)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
obtaining, at a computing device, user behavior data in connection with a user of an application;
detecting, by the computing device, an action by another user related to an interaction with the application;
generating, by the computing device, user behavior feature data based on the detected action by the other user using the obtained user behavior data;
determining, by the computing device, via execution of multiple anomaly prediction models, a plurality of user behavior anomaly predictions, each user behavior anomaly prediction being a different type of anomaly prediction, each of the user behavior anomaly predictions indicating a probability that user behavior represented by the user behavior feature data is anomalous user behavior;
aggregating, by the computing device, the different type of the plurality of user behavior anomaly predictions from the multiple anomaly prediction models, the aggregation comprising applying a weight value to a portion of the different type of the plurality of user behavior anomaly predictions;
determining, by the computing device, an aggregate user anomaly behavior prediction based on the aggregating of the different type of the plurality of user behavior anomaly predictions;
making, by the computing device, a user behavior anomaly determination using the aggregate user behavior anomaly prediction, the user behavior anomaly determination indicating whether or not the user behavior represented by the user behavior feature data is anomalous user behavior; and
providing, by the computing device, functionality for the application to control the action based on the user behavior anomaly determination.