System and method for identifying related credit inquiries

• US 10,592,982 B2
• Filed: 02/16/2017
• Issued: 03/17/2020
• Est. Priority Date: 03/14/2013
• Status: Active Grant

• Current Assignee: Csidentity Corporation (Experian plc)
• Original Assignee: Csidentity Corporation (Experian plc)
• Inventors: Samler, Steven E., Hall, Matthew
• Primary Examiner(s): Wong, Eric T
• Application Number:
  US15/434,551
• Publication Number:
  US 20170287065A1
• Time in Patent Office: 1,125 Days
• CPC Class Codes: G06Q 40/03 Credit; Loans; Processing thereof

Abstract

In one embodiment, a method includes receiving a credit inquiry for a monitored consumer. The method further includes performing at least one credit-inquiry analysis on the credit inquiry. The performing yields at least one of a matched creditor from master creditor records and an inferred credit category based on a lexical analysis. In addition, the method includes, responsive to the performing, determining whether the credit inquiry is related to a previous credit inquiry for the monitored consumer. Further, the method includes, responsive to a determination that the credit inquiry is related to a previous credit inquiry, suppressing a credit alert to the monitored consumer.

Claims

1. A system comprising:

one or more databases configured to electronic files of consumers and information related to credit inquiries on the credit files of the consumers;

a server computer communicably coupled to the one or more databases, wherein the server computer is programmed to;

access an electronic file of a monitored user to identify a credit event, wherein the electronic file is stored in a data store;

extract, from the electronic file information related to the credit event, the extracted information comprising a name of an initiator of the credit event and at least one other field;

perform name matching using the name of the initiator of the credit event and stored names in master records;

perform token-based matching using the at least one other field and corresponding fields in the master records;

perform a lexical analysis of the name of the initiator of the inquiry event;

determine an agreed-upon multi-module match among the name matching, the token-based matching, and the lexical analysis;

determine, based at least in part on the agreed-upon multi-module match whether there is a related credit event of the monitored user that occurred within a predetermined window of time of the identified credit event; and

responsive to a determination that there is not a related credit event of the monitored user that occurred within a predetermined window of time of the identified credit event, generate and send an alert of the identified credit event to the monitored user.

2. The system of claim 1, the one or more databases is further configured to store master creditor records which comprise a master name list and a master address list.

3. The system of claim 2, performing the name matching analysis comprises looking up the name of the initiator from the master name list and returning creditors from the master records whose name matches the name of the initiator.

4. The system of claim 2, performing the token based matching comprises:

creating a source token set based on an address of the initiator;

creating a target token set based on an address of a creditor in a master address list;

categorizing each token in the source token set and each token in the target set into a plurality of categories;

filtering the source token set and the target token set to remove tokens deemed insignificant;

for each of the source token set and the target token set, combining sequential tokens that have been categorized a same category;

identifying each instance of an exact match between a token of the source token set and a token of the target token set;

removing the token of the source token set from the source token set and removing the token of the target token set from the target token set where the exact match occur; and

detecting a match between the source token set and the target set in response to a determination that the source token set and the target token set are empty.

5. The system of claim 2, the server computer is further programmed to perform a thesaurus lookup on the name of the initiator to identify a match from the master records.

6. The system of claim 1, the lexical analysis comprises an n-grams analysis on a sequence of words in the name of the initiator.

7. The system of claim 1, the server computer is further programmed to:

responsive to a determination that there is a related credit event of the monitored user that occurred within the predetermined window of time of the identified credit event, preventing the computer system from generating the alert.

8. A method comprising, by a computer system comprising at least one processor and memory:

accessing an electronic file of a monitored user to identify a credit event, wherein the electronic file is stored in a data store;

extracting, from the electronic file information related to the credit event, the extracted information comprising a name of an initiator of the credit event and at least one other field;

performing name matching using the name of the initiator of the credit event and stored names in master records;

performing token-based matching using the at least one other field and corresponding fields in the master records;

performing a lexical analysis of the name of the initiator of the inquiry event;

determining an agreed-upon multi-module match among the name matching, the token-based matching, and the lexical analysis;

determining, based at least in part on the agreed-upon multi-module match whether there is a related credit event of the monitored user that occurred within a predetermined window of time of the identified credit event; and

responsive to a determination that there is not a related credit event of the monitored user that occurred within a predetermined window of time of the identified credit event, generating and sending an alert of the identified credit event to the monitored user.

9. The method of claim 8, performing the name matching analysis comprises looking up the name of the initiator from the master records and returning creditors from the master records whose name matches the name of the initiator.

10. The method of claim 8, performing the token-based matching comprises:

creating a source token set based on an address of the initiator;

creating a target token set based on an address of a creditor in a master address list;

categorizing each token in the source token set and each token in the target set into a plurality of categories;

filtering the source token set and the target token set to remove tokens deemed insignificant;

for each of the source token set and the target token set, combining sequential tokens that have been categorized a same category;

identifying each instance of an exact match between a token of the source token set and a token of the target token set;

removing the token of the source token set from the source token set and removing the token of the target token set from the target token set where the exact match occur; and

detecting a match between the source token set and the target set in response to a determination that the source token set and the target token set are empty.

11. The method of claim 8, further comprising performing a thesaurus lookup on the name of the initiator to identify a match from the master records.

12. The method of claim 8, the lexical analysis comprises an n-grams analysis on a sequence of words in the name of the initiator.

13. The method of claim 8, further comprising:

responsive to a determination that there is a related credit event of the monitored user that occurred within the predetermined window of time of the identified credit event, preventing the computer system from generating the alert.

14. A non-transitory computer readable medium storing software instructions thereon, the software instructions cause a computer system to:

access an electronic file of a monitored user to identify a credit event, wherein the electronic file is stored in a data store;

extract, from the electronic file information related to the credit event, the extracted information comprising a name of an initiator of the credit event and at least one other field;

perform name matching using the name of the initiator of the credit event and stored names in master records;

perform token-based matching using the at least one other field and corresponding fields in the master records;

perform a lexical analysis of the name of the initiator of the inquiry event;

determine an agreed-upon multi-module match among the name matching, the token-based matching, and the lexical analysis;

determine, based at least in part on the agreed-upon multi-module match whether there is a related credit event of the monitored user that occurred within a predetermined window of time of the identified credit event; and

responsive to a determination that there is not a related credit event of the monitored user that occurred within a predetermined window of time of the identified credit event, generate and send an alert of the identified credit event to the monitored user.

15. The non-transitory computer readable medium of claim 14, performing the name matching analysis comprises looking up the name of the initiator from the master records and returning creditors from the master records whose name matches the name of the initiator.

16. The non-transitory computer readable medium of claim 15, the software instructions further cause the computer system to perform a thesaurus lookup on the name of the initiator to identify a match from the master records.

17. The non-transitory computer readable medium of claim 15, the master records comprises a master name list and a master address list.

18. The non-transitory computer readable medium of claim 14, performing the token based matching comprises:

creating a source token set based on an address of the initiator;

creating a target token set based on an address of a creditor in a master address list;

categorizing each token in the source token set and each token in the target set into a plurality of categories;

filtering the source token set and the target token set to remove tokens deemed insignificant;

for each of the source token set and the target token set, combining sequential tokens that have been categorized a same category;

identifying each instance of an exact match between a token of the source token set and a token of the target token set;

removing the token of the source token set from the source token set and removing the token of the target token set from the target token set where the exact match occur; and

detecting a match between the source token set and the target set in response to a determination that the source token set and the target token set are empty.

19. The non-transitory computer readable medium of claim 14, the lexical analysis comprises an n-grams analysis on a sequence of words in the name of the initiator.

20. The non-transitory computer readable medium of claim 14, the software instructions further cause the computer system to:

responsive to a determination that there is a related credit event of the monitored user that occurred within the predetermined window of time of the identified credit event, prevent the computer system from generating the alert.