Detecting relationships across data columns
AssigneeOptum Technology, Inc.
InventorsSwapna Sourav Rout, Sruti Rallapalli, Mrinalini M, Kirk Michael Wroblewski, Bhuvaneshwari Atti Janakiram
Life of this patent
Estimating the term…
Abstract
There is a need for more effective and efficient detection of cross-data-column relationships. This need can be addressed by, for example, techniques for detecting cross-data-column data relationships that utilize at least one of feature-based similarity models and deep-learning-based similarity models. The cross-data-column data relationships may be displayed to an end-user using a cross-column relationship detection user interface.
Claims20
A computer-implemented method comprising:
- inputting, by one or more processors and to a trained feature-based similarity model, (1) a plurality of input data columns comprising an untagged data column associated with a first database, and (2) one or more tagged data columns associated with a second database;
- obtaining, by the one or more processors and from the trained feature-based similarity model, a related subset of the one or more tagged data columns for the untagged data column, wherein the trained feature-based similarity model is configured to:
- (a) for a particular tagged data column of the one or more tagged data columns:
- (i) determine one or more feature-based similarity measures;
- and
- (ii) combine the one or more feature-based similarity measures to determine a weighted similarity score set comprising one or more weighted similarity scores for the particular tagged data column;
- and
- (b) determine the related subset based at least in part on the weighted similarity score set;
- and
- generating, by the one or more processors and based at least in part on the related subset and one or more database consolidation operations with respect to the first database and the second database, a consolidated database.
A system comprising one or more processors and at least one memory storing processor-executable instructions that, when executed by the one or more processors, cause the system to:
- input, to a trained feature-based similarity model, (1) a plurality of input data columns comprising an untagged data column associated with a first database, and (2) one or more tagged data columns associated with a second database;
- obtain, from the trained feature-based similarity model, a related subset of the one or more tagged data columns for the untagged data column, wherein the trained feature-based similarity model is configured to:
- (a) for a particular tagged data column of the one or more tagged data columns:
- (i) determine one or more feature-based similarity measures;
- and
- (ii) combine the one or more feature-based similarity measures to determine a weighted similarity score set comprising one or more weighted similarity scores for the particular tagged data column;
- and
- (b) determine the related subset based at least in part on the weighted similarity score set;
- and
- generate, based at least in part on the related subset and one or more database consolidation operations with respect to the first database and the second database, a consolidated database.
One or more non-transitory computer-readable storage media comprising instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
- inputting, to a trained feature-based similarity model, (1) a plurality of input data columns comprising an untagged data column associated with a first database, and (2) one or more tagged data columns associated with a second database;
- obtaining, from the trained feature-based similarity model, a related subset of the one or more tagged data columns for the untagged data column, wherein the trained feature-based similarity model is configured to:
- (a) for a particular tagged data column of the one or more tagged data columns:
- (i) determine one or more feature-based similarity measures;
- and
- (ii) combine the one or more feature-based similarity measures to determine a weighted similarity score set comprising one or more weighted similarity scores for the particular tagged data column;
- and
- (b) determine the related subset based at least in part on the weighted similarity score set;
- and
- generate, based at least in part on the related subset and one or more database consolidation operations with respect to the first database and the second database, a consolidated database.
Dependencies are inferred from the claim text.