Granted Patent B2

US 12,265,950 · App. 17/564,363 · Granted Apr 1, 2025

Inventors: Jingshuai Zhang (Beijing, CN); Qifan Huang (Beijing, CN); Chao Ma (Beijing, CN); Hengshu Zhu (Beijing, CN); Peng Wang (Beijing, CN); Kaichun Yao (Beijing, CN); Jing Wang (Beijing, CN)

Assignee: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.

G06Q10/1053G06N3/08

Abstract

Claims (67)

1. A method of matching data, comprising:

obtaining a first instance of a resume and a second instance of a job profile;

determining, for a meta path, a resume feature data of the first instance and a profile feature data of the second instance, wherein the meta path is a knowledge graph path from the resume to the job profile;

applying a classification model to the resume feature data of the first instance and the profile feature data of the second instance to determine a matching result between the first instance and the second instance;

obtaining, from a set of meta paths comprising the meta path, at least one meta path related to a selection factor for the second instance, wherein the selection factor is one of a plurality of reasons for matching the first instance with the second instance;

determining a score of the selection factor based on resume feature information and profile feature information for the at least one meta path, so as to perform a ranking on the plurality of reasons, wherein each meta path includes a head entity, a tail entity, and at least one intermediate entity, a type of the head entity is a resume type, a type of the tail entity is a job profile type, a type of the at least one intermediate entity includes at least one of a field type, a skill type, and a job title type, and meta paths corresponding to different reasons have different combinations of entities and relations;

determining a target type in a target meta path in the at least one meta path;

determining an indication value indicating an importance of an instance of the target type based on a similarity between the first instance and the instance of the target type and a similarity between the second instance and the instance of the target type; and

displaying the instance of the target type in association with the selection factor based on the indication value, wherein

a classifier is trained by using a plurality of sample resume instances, a plurality of sample job profile instances and a sample matching result between the plurality of sample resume instances and the plurality of sample job profile instances.

2. The method of claim 1 , wherein determining the resume feature data and the profile feature data comprises:

determining a number of instances of the meta path from the first instance to the second instance;

determining a similarity of the first instance and the second instance for the meta path based on the number of instances of the meta path; and

determining the resume feature data and the profile feature data based on the similarity.

3. The method of claim 1 , wherein determining the resume feature data and the profile feature data comprises:

obtaining a first plurality of reference resume feature data corresponding to a first plurality of reference instances of the resume;

determining whether there is a reference instance matching the first instance among the first plurality of reference instances or not based on an entity type instance in the meta path included in the first instance; and

using, in response to there being a reference instance corresponding to the first instance among the first plurality of reference instances, a reference resume feature data corresponding to the reference instance as the resume feature data.

4. The method of claim 3 , wherein the determining whether there is a reference instance matching the first instance among the first plurality of reference instances or not based on an entity type instance in the meta path included in the first instance comprises:

determining whether there is a reference instance corresponding to the first instance or not according to a number or a percentage of the entity type instance in the meta path included in the first instance matched with the entity type instance in the reference instance.

5. The method of claim 3 , wherein determining the resume feature data and the profile feature data further comprises:

obtaining a second plurality of reference profile feature data corresponding to a second plurality of reference instances of the job profile;

determining whether there is a reference instance corresponding to the second instance among the second plurality of reference instances or not based on an entity type instance in the meta path included in the second instance; and

using, in response to there being a reference instance corresponding to the second instance among the second plurality of reference instances, a reference profile feature data corresponding to the reference instance as the profile feature data.

6. The method of claim 1 , wherein determining the resume feature data and the profile feature data further comprises:

obtaining a second plurality of reference profile feature data corresponding to a second plurality of reference instances of the job profile;

determining whether there is a reference instance corresponding to the second instance among the second plurality of reference instances or not based on an entity type instance in the meta path included in the second instance; and

using, in response to there being a reference instance corresponding to the second instance among the second plurality of reference instances, a reference profile feature data corresponding to the reference instance as the profile feature data.

7. The method of claim 1 , wherein the classification model is at least one of: a logistic regression model, a neural network model and a gradient boosting decision tree model.

8. An electronic device, comprising:

at least one processor; and

a memory communicatively connected with the at least one processor; wherein,

the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor implements operations of matching data, comprising:

obtaining a first instance of a resume and a second instance of a job profile;

determining, for a meta path, a resume feature data of the first instance and a profile feature data of the second instance, wherein the meta path is a knowledge graph path from the resume to the job profile;

applying a classification model to the resume feature data of the first instance and the profile feature data of the second instance to determine a matching result between the first instance and the second instance;

obtaining, from a set of meta paths comprising the meta path, at least one meta path related to a selection factor for the second instance, wherein the selection factor is one of a plurality of reasons for matching the first instance with the second instance;

determining a score of the selection factor based on resume feature information and profile feature information for the at least one meta path, so as to perform a ranking on the plurality of reasons, wherein each meta path includes a head entity, a tail entity, and at least one intermediate entity, a type of the head entity is a resume type, a type of the tail entity is a job profile type, a type of the at least one intermediate entity includes at least one of a field type, a skill type, and a job title type, and meta paths corresponding to different reasons have different combinations of entities and relations;

determining a target type in a target meta path in the at least one meta path;

determining an indication value indicating an importance of an instance of the target type based on a similarity between the first instance and the instance of the target type and a similarity between the second instance and the instance of the target type; and

displaying the instance of the target type in association with the selection factor based on the indication value, wherein

a classifier is trained by using a plurality of sample resume instances, a plurality of sample job profile instances and a sample matching result between the plurality of sample resume instances and the plurality of sample job profile instances.

9. The electronic device of claim 8 , wherein the at least one processor further implements:

determining a number of instances of the meta path from the first instance to the second instance;

determining a similarity of the first instance and the second instance for the meta path based on the number of instances of the meta path; and

determining the resume feature data and the profile feature data based on the similarity.

10. The electronic device of claim 8 , wherein the at least one processor further implements:

obtaining a first plurality of reference resume feature data corresponding to a first plurality of reference instances of the resume;

determining whether there is a reference instance matching the first instance among the first plurality of reference instances or not based on an entity type instance in the meta path included in the first instance; and

using, in response to there being a reference instance corresponding to the first instance among the first plurality of reference instances, a reference resume feature data corresponding to the reference instance as the resume feature data.

11. The electronic device of claim 10 , wherein the at least one processor further implements:

determining whether there is a reference instance corresponding to the first instance or not according to a number or a percentage of the entity type instance in the meta path included in the first instance matched with the entity type instance in the reference instance.

12. The electronic device of claim 10 , wherein the at least one processor further implements:

obtaining a second plurality of reference profile feature data corresponding to a second plurality of reference instances of the job profile;

determining whether there is a reference instance corresponding to the second instance among the second plurality of reference instances or not based on an entity type instance in the meta path included in the second instance; and

using, in response to there being a reference instance corresponding to the second instance among the second plurality of reference instances, a reference profile feature data corresponding to the reference instance as the profile feature data.

13. The electronic device of claim 8 , wherein the classification model is at least one of: a logistic regression model, a neural network model and a gradient boosting decision tree model.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are configured to cause the computer implement operations of matching data, comprising:

obtaining a first instance of a resume and a second instance of a job profile;

determining, for a meta path, a resume feature data of the first instance and a profile feature data of the second instance, wherein the meta path is a knowledge graph path from the resume to the job profile;

applying a classification model to the resume feature data of the first instance and the profile feature data of the second instance to determine a matching result between the first instance and the second instance;

obtaining, from a set of meta paths comprising the meta path, at least one meta path related to a selection factor for the second instance, wherein the selection factor is one of a plurality of reasons for matching the first instance with the second instance;

determining a score of the selection factor based on resume feature information and profile feature information for the at least one meta path, so as to perform a ranking on the plurality of reasons, wherein each meta path includes a head entity, a tail entity, and at least one intermediate entity, a type of the head entity is a resume type, a type of the tail entity is a job profile type, a type of the at least one intermediate entity includes at least one of a field type, a skill type, and a job title type, and meta paths corresponding to different reasons have different combinations of entities and relations;

determining a target type in a target meta path in the at least one meta path;

determining an indication value indicating an importance of an instance of the target type based on a similarity between the first instance and the instance of the target type and a similarity between the second instance and the instance of the target type; and

displaying the instance of the target type in association with the selection factor based on the indication value, wherein

a classifier is trained by using a plurality of sample resume instances, a plurality of sample job profile instances and a sample matching result between the plurality of sample resume instances and the plurality of sample job profile instances.

Assignments (2)

Priority Claims (1)

Continuity (1)

References Cited (14)