Cognitive data anonymization
A computer implemented method for data anonymization comprises: receiving a request for data that needs anonymization. The request comprises at least one field descriptor of data to be retrieved and a usage scenario of a user for the requested data. Then, based on the usage scenario, an anonymization algorithm to be applied to the data that is referred to by the field descriptor is determined. Subsequently, the determined anonymization algorithm is applied to the data that is referred to by the field descriptor. A testing is performed, as to whether the degree of anonymization fulfills a requirement that is related to the usage scenario. In the case, the requirement is fulfilled, access to the anonymized data is provided.
1. A computer program product for anonymizing data, the computer program comprising:
a non-transitory computer readable storage medium having computer usable code embodied therewith, the computer usable program code comprising:
computer usable code configured for receiving a request for data, the request comprising at least one field descriptor of data to be retrieved and a usage scenario of a user for the requested data;
computer usable code configured for determining, based on the usage scenario an anonymization algorithm to be applied to the data that is referred to by the field descriptor;
computer usable code configured for applying the anonymization algorithm to the data that is referred to by the field descriptor;
computer usable code configured for testing, whether a degree of anonymization fulfills a requirement that is related to the usage scenario;
computer usable code configured for, based on a determination that the requirement is fulfilled, providing access to the anonymized data, wherein the usage scenario comprises an intent of the user requesting the data and a role of the user.
2. The computer program product of claim 1 , wherein at least one protection level is associated with the field descriptor.
3. The computer program product of claim 2 , wherein the anonymization algorithm to be applied is selected based on the usage scenario and the protection level.
4. The computer program product of claim 1 , wherein the anonymization algorithm is at least one of the following: generalization, redaction, suppressing, sampling, randomization, data swapping, masking, enumeration.
5. The computer program product of claim 1 , wherein the computer usable program code further comprises computer usable code configured for receiving a feedback from a user, the feedback forming a support of machine learning.
6. The computer program product of claim 1 , wherein the requirement that is related to the usage scenario is k-anonymity.
7. The computer program product of claim 1 , wherein the data is related to information asset metadata that comprises an availability of data tables to the user and the field descriptors these data tables have.
8. The computer program product of claim 1 , wherein business terms are provided, forming keywords to be used by the user, wherein classification information is provided, representing relationships between business terms and information asset metadata.
9. The computer program product of claim 1 , wherein generalization hierarchies are provided for enabling a generalization algorithm to execute.
10. The computer program product of claim 1 , wherein a data privacy governance rule and a corresponding test algorithm for the testing, is specified, for controlling the method to be executed until, based on the data to be retrieved and the usage scenario, the retrieved data after anonymization corresponds to the degree of anonymization.
11. The computer program product of claim 1 , wherein a model is provided, the model specifying a relationship between an anonymization algorithm, a usage scenario, and a classification of the data.
12. The computer program product of claim 11 , wherein the model is built up by machine learning.
13. The computer program product of claim 12 , wherein the model is adapted, by machine learning, during its usage.
14. The computer program product of claim 1 , wherein an enumeration for a specification of the usage scenario is provided.
15. The computer program product of claim 14 , wherein the enumeration may specify, at least one of medical study evaluation, customer churn analysis, fraud detection.
16. The computer program product of claim 1 , wherein the computer usable program code further comprises a catalog comprising business terms, information asset metadata, classification information, a generalization hierarchy, a data privacy governance rule and a corresponding test algorithm, an enumeration specifying the usage scenario.
17. The computer program product of claim 1 , wherein an interaction with the user is admitted so that the user may, using machine learning, amend the method of anonymization.
18. A system for anonymizing data, the system comprising:
a communication component embodied at least partially in hardware and configured to communicate with a user;
a processor, communicatively coupled to the communication component, the processor comprising:
a means for receiving a request for data, the request comprising at least one field descriptor of data to be retrieved and a usage scenario of a user for the requested data;
a means for determining, based on the usage scenario, an anonymization algorithm to be applied to the data that is referred to by the field descriptor;
a means for applying the determined anonymization algorithm to the data that is referred to by the field descriptor;
a means for testing, whether a degree of anonymization fulfills a requirement that is related to the usage scenario;
a means for, providing access to the anonymized data based on a determination that the requirement is fulfilled, wherein the usage scenario comprises an intent of the user requesting the data and a role of the user.