Verification of Data Categorization
AssigneeInternational Business Machines Corporation
InventorLeho Nigul
Life of this patent
Estimating the term…
Abstract
Verification and categorization of data in a system that interfaces with common knowledge repositories having different application programming interfaces. The system inputs a data tree structure with categories of information. The relationships in the data tree are queried against common knowledge repositories. A report of potentially erroneous categorizations in the data tree may be output for further review.
Claims20
A computer-implemented method of verifying a data categorization in a categorization verification system, the method including executable instructions that perform actions comprising:
- determining, within a framework that supports queries, an accuracy of relationships between data that is categorized in a data structure based on information received from a common knowledge repository;
- responsive to a determination of an incorrect categorization, flagging a relationship within the data structure as potentially inaccurate;
- and
- reporting the potentially inaccurate categorization.
A system that verifies data categorization, the system comprising:
- a verification engine that determines an accuracy of a categorization of input data in a data structure;
- an interface communicatively connected to the verification engine;
- a common knowledge repository configured to accept a communication from the interface;
- and
- a plug-in framework that provides connectively supports the interface and the repository.
A computer program product comprising computer recordable media comprising program code, which when executed by a processor, perform actions comprising:
- determining, within a framework that supports queries to a common knowledge repository, an accuracy of relationships between data that is categorized in a data structure based on information received from a common knowledge repository;
- responsive to a determination of an incorrect categorization, flagging a relationship within the data structure as potentially inaccurate;
- and
- reporting the potentially inaccurate categorization.
Dependencies are inferred from the claim text.