Data matching accuracy based on context features

Patent metadata

Field	Value
Publication number	US-11074230-B2
Application number	US-201816120949-A
Country	US
Kind code	B2
Filing date	Sep 4, 2018
Priority date	Sep 4, 2018
Publication date	Jul 27, 2021
Grant date	Jul 27, 2021

Abstract

A method is provided for data matching between a set of source data structures and a set of target data structures. The method includes obtaining, using a processor device configured to perform machine learning, source to target matching results with matching scores, based on the sets of source and target data structures. The method further includes calculating, by the processor device, context information for data structure pairs based on a structure similarity and an ontology similarity between constituent data structures thereof. Each of data structure pairs include as the constituent data structures a respective source data structure and a respective target data structure from the sets of source and target data structures. The method also includes updating, by the processor device, the matching scores based on the context information. The method additionally includes controlling, by the processor device, a hardware device responsive to at least one updated matching score.

First claim

What is claimed is:

1. A computer-implemented method for data matching between a set of source data structures and a set of target data structures, the method comprising:
obtaining, using a processor device configured to perform machine learning, source to target matching results with matching scores, based on the sets of source and target data structures;
calculating, by the processor device, context information for data structure pairs based on a structure similarity and an ontology similarity between constituent data structures thereof, each of data structure pairs comprising as the constituent data structures a respective source data structure and a respective target data structure from the sets of source and target data structures;
updating, by the processor device, the matching scores based on the context information;
controlling, by the processor device, a hardware device responsive to at least one of the updated matching scores; and
wherein the method further comprises determining the structure similarity by finding all neighborhood mappings from an initial mapping and computing an aggregation score for each of the data structure pairs, and wherein the matching scores for the data structure pairs having the aggregation score greater than a threshold are increased by said updating step.

2. The computer-implemented method of claim 1 , wherein for a relational database, the context information based on the structure similarity is, in turn, based on a respective source table and a respective target table and columns in the respective source and target tables.

3. The computer-implemented method of claim 2 , wherein the structure similarity score for a given data structure pair increases responsive to most of the columns matching between the respective source and target tables that comprise the given data structure pair, and wherein the structure similarity score for a column pair formed from a respective column in each of the respective source and target tables increases responsive to the structure similarity score for the respective source and target tables that include the column pair being greater than a threshold.

4. The computer-implemented method of claim 2 , wherein the context information based on the source structure is calculated, in turn, based on a first premise that the respective source and target tables are considered to match responsive to most of the columns therebetween matching, and a second premise that at least one column in each of the respective source and target tables are considered to match responsive to the respective source and target tables being considered a match.

5. The computer-implemented method of claim 1 , wherein for extensible markup language, the context information based on the structure similarity is, in turn, based on layered elements and attributes of the layered elements.

6. The computer-implemented method of claim 1 , wherein the context information based on the ontology structure is calculated, in turn, based on a premise that a particular element of a data structure is likely to match an ontology target responsive to a neighboring element of the particular element matching a related ontology target with respect to the ontology target.

7. The computer-implemented method of claim 1 , further comprising determining the ontology similarity by finding all neighborhood mappings having an ontology relationship from an initial mapping and computing an ontology distance between members of the data structure pairs, wherein the matching scores for the data structure pairs having the ontology distance greater than a threshold are increased by said updating step.

8. The computer-implemented method of claim 1 , wherein the method is performed by a server comprised in a cloud computing platform, and the processor device is comprised in the server.

9. A computer program product for data matching between a set of source data structures and a set of target data structures, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform a method comprising:
obtaining, using a processor device configured to perform machine learning, source to target matching results with matching scores, based on the sets of source and target data structures;
calculating, by the processor device, context information for data structure pairs based on a structure similarity and an ontology similarity between constituent data structures thereof, each of data structure pairs comprising as the constituent data structures a respective source data structure and a respective target data structure from the sets of source and target data structures;
updating, by the processor device, the matching scores based on the context information;
controlling, by the processor device, a hardware device responsive to at least one of the updated matching scores; and
wherein the method further comprises determining the structure similarity by finding all neighborhood mappings from an initial mapping and computing an aggregation score for each of the data structure pairs, and wherein the matching scores for the data structure pairs having the aggregation score greater than a threshold are increased by said updating step.

10. The computer program product of claim 9 , wherein for a relational database, the context information based on the structure similarity is, in turn, based on a respective source table and a respective target table and columns in the respective source and target tables.

11. The computer program product of claim 10 , wherein the structure similarity score for a given data structure pair increases responsive to most of the columns matching between the respective source and target tables that comprise the given data structure pair, and wherein the structure similarity score for a column pair formed from a respective column in each of the respective source and target tables increases responsive to the structure similarity score for the respective source and target tables that include the column pair being greater than a threshold.

12. The computer program product of claim 10 , wherein the context information based on the source structure is calculated, in turn, based on a first premise that the respective source and target tables are considered to match responsive to most of the columns therebetween matching, and a second premise that at least one column in each of the respective source and target tables are considered to match responsive to the respective source and target tables being considered a match.

13. The computer program product of claim 9 , wherein for extensible markup language, the context information based on the structure similarity is, in turn, based on layered elements and attributes of the layered elements.

14. The computer program product of claim 9 , wherein the context information based on the ontology structure is calculated, in turn, based on a premise that a particular element of a data structure is likely to match an ontology target responsive to a neighboring element of the particular element matching a related ontology target with respect to the ontology target.

15. The computer program product of claim 9 , wherein the method further comprises determining the ontology similarity by finding all neighborhood mappings having an ontology relationship from an initial mapping and computing an ontology distance between members of the data structure pairs, wherein the matching scores for the data structure pairs having the ontology distance greater t

Assignees

• IBM

Inventors

• Liu Pei Ni
• Jiang Jian Min
• Xu Jing Min
• Ruan Yaoping
• Lei Hui
• Smith Michael
• Crawford Bradley W

Classifications

• G06F16/211 Primary

  Schema design and management · CPC title

• G06F16/254 Primary

  Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses · CPC title