Intelligent data enrichment using knowledge graph

Patent metadata

Field	Value
Publication number	US-2022237185-A1
Application number	US-202117160153-A
Country	US
Kind code	A1
Filing date	Jan 27, 2021
Priority date	Jan 27, 2021
Publication date	Jul 28, 2022
Grant date	—

Abstract

A computer-implemented method can receive a source table containing data tuples and a source schema defining attributes of the data tuples, and match the source schema to an ontology of a knowledge graph. The knowledge graph can include a plurality of instances and the ontology defines properties of the plurality of instances. The computer-implemented method can link the data tuples to respective instances in the knowledge graph, and identifying non-matching properties of the respective instances, wherein the non-matching properties are defined in the ontology and not matched to the source schema. The computer-implemented method can obtain property values associated with the non-matching properties from the knowledge graph, and add one or more of the non-matching properties and the associated property values to respective data tuples of the source table.

First claim

1 . A computer-implemented method, implemented by one or more computing devices comprising at least one hardware processor and one or more tangible memories coupled to the at least one hardware processor, the method comprising:
receiving a source table containing data tuples and a source schema defining attributes of the data tuples; matching the source schema to an ontology of a knowledge graph, wherein the knowledge graph comprises a plurality of instances and the ontology defines properties of the plurality of instances; linking the data tuples to respective instances in the knowledge graph; identifying non-matching properties of the respective instances, wherein the non-matching properties are defined in the ontology and not matched to the source schema; obtaining property values associated with the non-matching properties from the knowledge graph; and adding one or more of the non-matching properties and the associated property values to respective data tuples of the source table.

2 . The method of claim 1 , further comprising creating a ranked list of non-matching properties and the associated property values, wherein the creating comprises ranking the non-matching properties of the respective instances.

3 . The method of claim 2 , further comprising presenting the ranked list of non-matching properties and the associated property values on a user interface, and receiving an input from the user interface indicating which non-matching properties and the associated properties values are added to respective data tuples of the source table.

4 . The method of claim 2 , wherein ranking the non-matching properties of the respective instances comprises counting instances in the knowledge graph having property values associated with the non-matching properties.

5 . The method of claim 1 , further comprising logging data linkage information, wherein the data linkage information records at least which non-matching properties have been added to respective data tuples of the source table.

6 . The method of claim 5 , further comprising ranking the non-matching properties of the respective instances based on previously logged data linkage information.

7 . The method of claim 1 , further comprising assigning a null value to a data tuple of the source table if a corresponding non-matching property of the respective instance linked to the data tuple has no associated property value or if the data tuple is not linked to a respective instance in the knowledge graph.

8 . The method of claim 7 , further comprising ranking the non-matching properties of the respective instances based on minimizing a total number of null values assigned to the data tuple of the source table.

9 . The method of claim 1 , further comprising mapping the data tuples to respective property tuple lists, wherein a property tuple list corresponds to a respective instance linked to the data tuple and comprises one or more pairs of non-matching property and property value associated with the non-matching property.

10 . The method of claim 1 , wherein the knowledge graph is one of a plurality of knowledge graphs, wherein the method further comprises identifying non-matching properties and obtaining property values associated with the non-matching properties from the plurality of knowledge graphs.

11 . A system comprising:
one or more processors; and memory coupled to the one or more processors comprising instructions causing the one or more processors implementing a table enrichment engine, wherein the table enrichment engine comprises: a user interface configured to receive a source table containing data tuples and a source schema defining attributes of the data tuples; a schema matching operator configured to match the source schema to an ontology of a knowledge graph, wherein the knowledge graph comprises a plurality of instances and the ontology defines properties of the plurality of instances; an instance linking operator configured to link the data tuples to respective instances in the knowledge graph; a new property finder configured to identify non-matching properties of the respective instances, wherein the non-matching properties are defined in the ontology and not matched to the source schema; a property value retriever configured to obtain property values associated with the non-matching properties from the knowledge graph; and an inserter configured to add one or more of the non-matching properties and the associated property values to respective data tuples of the source table.

12 . The system of claim 11 , wherein the table enrichment engine further comprises a ranking operator configured to rank the non-matching properties of the respective instances so as to create a ranked list of non-matching properties and the associated property values.

13 . The system of claim 12 , wherein the user interface is configured to present the ranked list of non-matching properties and the associated property values to a user, and receive an input from the user indicating which non-matching properties and the associated property values are added to respective data tuples of the source table.

14 . The system of claim 12 , wherein the ranking operator is configured to rank the non-matching properties of the respective instances based on counting instances in the knowledge graph having property values associated with the non-matching properties.

15 . The system of claim 12 , wherein the table enrichment engine further comprises a data logger configured to log data linkage information, wherein the data linkage information records at least which non-matching properties have been added to respective data tuples of the source table.

16 . The system of claim 15 , wherein the ranking operator is configured to rank the non-matching properties of the respective instances based on previously logged data linkage information.

17 . The system of claim 12 , wherein the inserter is configured to assign a null value to a data tuple of the source table if a corresponding non-matching property of the respective instance linked to the data tuple has no associated property value or if the data tuple is not linked to a respective instance in the knowledge graph.

18 . The system of claim 17 , wherein the ranking operator is configured to rank the non-matching properties of the respective instances based on minimizing a total number of null values assigned to the data tuple of the source table.

19 . The system of claim 11 , wherein the knowledge graph is one of a plurality of knowledge graphs, wherein the table enrichment engine is configured to identify non-matching properties and obtain property values associated with the non-matching properties from the plurality of knowledge graphs.

20 . One or more non-transitory computer-readable media having encoded thereon computer-executable instructions causing one or more processors to perform a method comprising:
receiving a source table containing data tuples and a source schema defining attributes of the data tuples; matching the source schema to an ontology of a knowledge graph, wherein the knowledge graph comprises a plurality of instances and the ontology defines properties of the plurality of instances; linking the data tuples to respective instances in the knowledge graph; identifying non-matching properties of the respective instances, wherein the non-matching properties are defined in the ontology and not matched to the source schema; obtaining property values associated with the non-matching propertie

Assignees

• Sap Se

Inventors

• Portisch Jan
• Bracholdt Sandra
• Huth Oliver
• Saggau Volker

Classifications

• G06F16/2465 Primary

  Query processing support for facilitating data mining operations in structured databases · CPC title

• G06F16/213 Primary

  with details for schema evolution support · CPC title