Systems and methods for classification of database records by mitigating bias using random data re-censoring
AssigneeCapital One Services, LLC
InventorsWendi LIU, Xue LI, Fei TONG, Bing LIU
Life of this patent
Estimating the term…
Abstract
Methods and systems are described herein for improving data processing efficiency of classifying user files in a database. More particularly, methods and systems are described herein for improving data processing efficiency of classifying user files in a database in which the user files have a temporal element. The methods and systems described herein accomplish these improvements by mitigating bias using random data censoring.
Claims20
A system for improving data processing efficiency of classifying user files in a database into cyber security states by mitigating bias using random data censoring, the system comprising:
- one or more processors;
- and
- a non-transitory, computer-readable medium comprising instructions that when executed by the one or more processors cause operations comprising:
- receiving a request to determine respective probabilities that a user record for a user belongs to each of a plurality of user record states, wherein the plurality of user record states comprise respective cyber security states;
- retrieving a user record history for the user record, wherein the user record history comprises periodic status checks, wherein each of the periodic status checks comprises a determination of a status of the user record;
- determining a set build date;
- determining a re-censoring probability based on a proximity of data in the user record history to the set build date;
- determining a respective user record status at each of the periodic status checks that is available based on the proximity;
- generating a re-censored record history based on the respective user record status at each of the periodic status checks that is available;
- processing the re-censored record history using a first model, wherein the first model comprises a survival analysis;
- receiving a first output of the first model indicating the respective probabilities that the user record for the user belongs to each of the plurality of user record states;
- and
- generating a cyber security recommendation based on the first output.
A method of improving data processing efficiency of classifying user files in a database into disparate states by mitigating bias using random data censoring, the method comprising:
- receiving a request to determine respective probabilities that a user record for a user belongs to each of a plurality of user record states;
- retrieving a user record history for the user record, wherein the user record history comprises periodic status checks;
- generating a re-censored record history by randomly re-censoring the user record history based on a set build date;
- generating a feature input for the user record history based on a respective user record status at each of the periodic status checks that is available in the re-censored record history;
- processing the feature input using a first model;
- receiving a first output of the first model indicating the respective probabilities that the user record for the user belongs to each of the plurality of user record states;
- and
- generating a recommendation based on the first output.
One or more non-transitory, computer-readable mediums comprising instructions that, when executed by one or more processors, cause operations comprising:
- receiving a request to determine respective probabilities that a user record for a user belongs to each of a plurality of user record states;
- retrieving a user record history for the user record, wherein the user record history comprises periodic status checks;
- generating a re-censored record history by re-censoring the user record history based on a set build date;
- processing the re-censored record history using a first model, wherein the first model comprises a survival analysis;
- receiving a first output of the first model indicating the respective probabilities that the user record for the user belongs to each of the plurality of user record states;
- and
- generating a cyber security recommendation based on the first output.
Dependencies are inferred from the claim text.