Inferring kinship based on graph data structure system and method thereof

Abstract

A system and method for inferring kinship relationships based on graph data structure. The method includes receiving data to be analyzed, which contains multiple customer records; generating multiple customer nodes based on the customer data, each customer node containing its own attribute information and one or more relationship attribute information; and executing a pattern matching algorithm to determine whether any two customer nodes satisfy a matching condition, and generating an edge indicating the association between the two customer nodes after determining that the matching condition is satisfied. By generating these customer nodes and edges, kinship relationships among the customers can be integrated.

Claims

1. A system for inferring kinship relationships based on graph data structure, comprising: a storage unit; and a processing unit electrically connected to the storage unit; wherein, The processing unit is configured to: receive data to be analyzed, the data containing multiple customer records; generate multiple customer nodes based on the customer data, each customer node containing its own attribute information and one or more relationship attribute information; execute a pattern matching algorithm, in which the pattern matching algorithm determines whether any two customer nodes satisfy a matching condition, and after determining that two customer nodes satisfy the matching condition, generates an edge indicating the association between the two customer nodes, wherein when the processing unit executes the pattern matching algorithm, after determining that two customer nodes satisfy the matching condition, it generates the edge indicating the association between the two customer nodes and having a connection attribute information based on the satisfied matching condition; the processing unit performs a logical verification based on the customer nodes and the edges with the connection attribute information between the customer nodes, wherein if the processing unit determines that the number of all edges where the customer node points to other customer nodes and has the connection attribute information of being a parent is greater than a threshold value, then the logical verification is deemed to fail; A customer relationship graph is generated based on the customer nodes and the edge, and the customer graph data is stored in the storage unit.

2. The system for inferring kinship based on graph data structure as described in claim 1, wherein, When the processing unit executes the pattern matching algorithm, the pattern matching algorithm includes a basic relation inference module and an extended relation inference module. When executing the basic relation inference module, the processing unit determines whether any two client nodes meet the matching condition based on their respective self-attribute information and / or relational attribute information. When executing the extended relation inference module, the processing unit determines whether any two client nodes meet the matching condition based on the client nodes and the edges generated when executing the basic relation inference module.

3. The system for inferring kinship based on graph data structure as described in claim 2, wherein, When the processing unit executes the basic relation inference module of the pattern matching algorithm, if the processing unit determines that at least two relation attribute information of one of the customer nodes respectively match at least two relation attribute information of another customer node, then it determines that the two customer nodes satisfy the matching condition.

4. The system for inferring kinship based on graph data structure as described in claim 2, wherein, When the processing unit executes the basic relation inference module of the pattern matching algorithm, if the processing unit determines that the relation attribute information of one of the customer nodes matches the self attribute information of another customer node, and another relation attribute information of one of the customer nodes matches the relation attribute information of another customer node, then it determines that the two customer nodes meet the matching condition.

5. The system for inferring kinship based on graph data structure as described in claim 2, wherein, When the processing unit executes the basic relation inference module of the pattern matching algorithm, the processing unit determines that the self-attribute information of one of the customer nodes matches the relation attribute information of another customer node, and the relation attribute information of one of the customer nodes matches the self-attribute information of another customer node, then it determines that the two customer nodes meet the matching condition.

6. The system for inferring kinship based on graph data structure as described in claim 2, wherein, When the processing unit executes the extended relationship inference module of the pattern matching algorithm, if the processing unit determines that one of the client nodes can be connected to another client node through other client nodes, then it determines that the two client nodes meet the matching condition.

7. The system for inferring kinship based on graph data structure as described in claim 1, wherein, After receiving multiple pending analysis data, the processing unit further determines whether the customer data in the pending analysis data are duplicates. If duplicates are found, the customer data in the pending analysis data with an older update date that is duplicated is excluded from the pending analysis data.

8. A method for inferring kinship relationships based on graph data structure, implemented by a processing unit, comprising: receiving data to be analyzed, the data containing multiple customer records; generating multiple customer nodes based on the customer data, each customer node containing its own attribute information and one or more relationship attribute information; executing a pattern matching algorithm, in which the pattern matching algorithm determines whether any two customer nodes satisfy a matching condition, and after determining that two customer nodes satisfy the matching condition, generating an edge indicating the association between the two customer nodes, wherein... When executing the pattern matching algorithm, after determining that the matching condition is met between two client nodes, an edge indicating the association between the two client nodes and having a connection attribute information is generated based on the met matching condition; and a logical verification is performed based on the client nodes and the edges with the connection attribute information between the client nodes, wherein if the processing unit determines that the number of all edges where the client node points to other client nodes and has the connection attribute information of being the parent is greater than a threshold value, then the logical verification is deemed to fail.

9. The method for inferring kinship based on graph data structure as described in claim 8, wherein, The pattern matching algorithm includes a basic relation inference module and an extended relation inference module. When executing the basic relation inference module, it determines whether any two client nodes meet the matching condition based on their own attribute information and / or their relation attribute information. When executing the extended relation inference module, it determines whether any two client nodes meet the matching condition based on the client nodes and the edges generated when executing the basic relation inference module.

10. The method for inferring kinship based on graph data structure as described in claim 9, wherein, When executing the basic relation inference module of the pattern matching algorithm, if it is determined that at least two relation attribute information of one of the client nodes respectively match at least two relation attribute information of another client node, then it is determined that the two client nodes satisfy the matching condition.

11. The method for inferring kinship based on graph data structure as described in claim 9, wherein, When executing the basic relation inference module of the pattern matching algorithm, if it is determined that the relation attribute information of one of the client nodes matches the self attribute information of another client node, and another relation attribute information of one of the client nodes matches the relation attribute information of another client node, then it is determined that the two client nodes meet the matching condition.

12. The method for inferring kinship based on graph data structure as described in claim 9, wherein, When executing the basic relation inference module of the pattern matching algorithm, if it is determined that the self-attribute information of one of the client nodes matches the relation attribute information of another client node, and the relation attribute information of one of the client nodes matches the self-attribute information of another client node, then it is determined that the two client nodes meet the matching condition.

13. The method for inferring kinship based on graph data structure as described in claim 9, wherein, When executing the extended relationship inference module of the pattern matching algorithm, if it is determined that one of the client nodes can be connected to another client node through other client nodes, then it is determined that the two client nodes meet the matching condition.

14. The method for inferring kinship based on graph data structure as described in claim 8, after receiving multiple pending analysis data, further includes: determining whether the customer data of the pending analysis data are duplicates of each other; if duplicates are found, excluding the customer data of the pending analysis data whose update date is older and which are duplicates from the pending analysis data.

Images