Entity resolution techniques for matching entity records from different data sources

Patent metadata

Field	Value
Publication number	US-11080272-B2
Application number	US-201916457666-A
Country	US
Kind code	B2
Filing date	Jun 28, 2019
Priority date	Jun 28, 2019
Publication date	Aug 3, 2021
Grant date	Aug 3, 2021

Abstract

Entity resolution techniques for matching entity records from different data sources are provided. In one technique, an entity record from a source database is identified along with multiple data items included therein. Each data item corresponds to an attribute of multiple source attributes. For one of the data items that corresponds to a first source attribute, multiple target attributes are identified. A first query is generated that includes the data items and associates the data item with each of the multiple target attributes. A second query that is different than the first query is also generated. Two searches are performed of a target database: one based on the first query and the other based on the second query. A scoring model generates multiple scores, one for each search result. It is determined whether the entity record matches an entity record in the target database based on the set of scores.

First claim

What is claimed is:

1. A method comprising:
identifying a first entity record from a source database;
identifying, in the first entity record, a plurality of data items, each of which corresponds to an attribute of a plurality of source attributes;
for a first data item, of the plurality of data items, that corresponds to a first source attribute of the plurality of source attributes, identifying a plurality of target attributes;
generating a first query that includes the plurality of data items and that associates the first data item with each of the plurality of target attributes;
generating a second query that is different than the first query;
performing a first search of a target database based on the first query, wherein the first search results in a first set of results;
performing a second search of the target database based on the second query, wherein the second search results in a second set of results;
using a scoring model to generate a set of scores, one score for each result in the first and second sets of results;
determining whether the first entity record matches an entity record in the target database based on the set of scores generated by the scoring model;
wherein the method is performed by one or more computing devices.

2. The method of claim 1 , further comprising:
for a second data item, of the plurality of data items, that corresponds to a second source attribute of the plurality of source attributes, identifying a second plurality of target attributes;
generating a first plurality of predicates, one for each target attribute of the second plurality of target attributes, wherein each predicate in the first plurality of predicates includes the second data item and is combined with each other predicate of first plurality of predicates with a disjunctive OR;
for a third data item, of the plurality of data items, that corresponds to a third source attribute of the plurality of source attributes, identifying a third plurality of target attributes;
generating a second plurality of predicates, one for each target attribute of the third plurality of target attributes, wherein each predicate in the second plurality of predicates includes the third data item and is combined with each other predicate of second plurality of predicates with the disjunctive OR;
wherein the second query includes the first plurality of predicates and the second plurality of predicates, wherein the first plurality of predicates and the second plurality of predicates are combined with the disjunctive OR.

3. The method of claim 1 , wherein:
generating the first query that includes generating a plurality of predicates, each predicate corresponding to a different data item-target attribute pair and including the first data item and a different target attribute of the plurality of target attributes;
the plurality of predicates are combined with a disjunctive OR.

4. The method of claim 1 , further comprising:
for a second data item, of the plurality of data items, that corresponds to a second source attribute of the plurality of source attributes, identifying a second plurality of target attributes;
wherein the first query associates the second data item with each of the second plurality of target attributes.

5. The method of claim 4 , wherein:
the first query includes a first compound predicate that associates the first data item with each of the plurality of target attributes and a second compound predicate that associates the second data item with each of the second plurality of target attributes;
the first compound predicate and the second compound predicate are combined using a conjunctive AND.

6. The method of claim 4 , wherein:
the plurality of target attributes is a first plurality of target attributes;
the first plurality of target attributes include first name and last name;
the first plurality of target attributes is the same as the second plurality of target attributes;
the first source attribute is first name and the second source attribute is last name.

7. The method of claim 1 , wherein:
the first source attribute is job title;
the plurality of target attributes include previous job title and current job title.

8. The method of claim 1 , wherein:
the first source attribute is organization name;
the plurality of target attributes include previous organization name and current organization name.

9. The method of claim 8 , further comprising:
for a second data item, of the plurality of data items, that corresponds to a second source attribute of the plurality of source attributes, identifying a second plurality of target attributes;
wherein the first query associates the second data item with each of the second plurality of target attributes;
wherein the second source attribute is job title;
the second plurality of target attributes include previous job title and current job title.

10. A method comprising:
identifying a first entity record from a source database;
identifying, in the first entity record, a plurality of data items, each of which corresponds to an attribute of a plurality of source attributes;
for a first data item, of the plurality of data items, that corresponds to a first source attribute of the plurality of source attributes, identifying a plurality of target attributes;
generating a first query that includes a plurality of predicates, each predicate corresponding to a different data item-target attribute pair and including the first data item and a different target attribute of the plurality of target attributes;
wherein the plurality of predicates are combined with a disjunctive OR;
performing a first search of a target database based on the first query, wherein the first search results in a first set of results;
using a scoring model to generate a set of scores, one for each result in the first set of results;
determining whether the first entity record matches an entity record in the target database based on the set of scores generated by the scoring model;
generating a second query that is different than the first query;
performing a second search of the target database based on the second query, wherein the second search results in a second set of results;
wherein using the scoring model comprises using the scoring model to generate a score for each result in the first and second sets of results;
wherein the method is performed by one or more computing devices.

11. One or more storage media storing instructions which, when executed by one or more processors, cause:
identifying a first entity record from a source database;
identifying, in the first entity record, a plurality of data items, each of which corresponds to an attribute of a plurality of source attributes;
for a first data item, of the plurality of data items, that corresponds to a first source attribute of the plurality of source attributes, identifying a plurality of target attributes;
generating a first query that includes the plurality of data items and that associates the first data item with each of the plurality of target attributes;
generating a second query that is different than the first query;
performing a first search of a target database based on the first query, wherein the first search results in a first set of results;
performing a second search of the target database based on the second query, wherein the second search results in a second set of results;
using a scoring model to generate a set of scores, one score for each result in the first and second sets of results;
determining whether the first entity record matches an entity record in the target database based on the set of scores generated by the scoring model

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Chen Yang
• Zhang Liang
• Zhao Haifeng
• Wang Jiashuo
• Krishnan Aparna
• Kishore Anand
• Wu Chencheng
• Moore John P

Classifications

• G06F16/2423 Primary

  Interactive query statement specification based on a database schema · CPC title

• G06F16/2425 Primary

  Iterative querying; Query formulation based on the results of a preceding query · CPC title

• G06F16/24578

  using ranking · CPC title

• G06F16/25

  Integrating or interfacing systems involving database management systems · CPC title