Systems and methods for providing recommendations based on collaborative and/or content-based nodal interrelationships

• US 10,423,880 B2
• Filed: 11/02/2015
• Issued: 09/24/2019
• Est. Priority Date: 09/28/2011
• Status: Active Grant

• Current Assignee: Nara Logics, Inc.
• Original Assignee: Nara Logics, Inc.
• Inventors: Wilson, Nathan R., Hueske, Emily A., Copeman, Thomas C.
• Primary Examiner(s): Hill, Stanley K.
• Application Number:
  US14/930,166
• Publication Number:
  US 20160055417A1
• Time in Patent Office: 1,422 Days
• CPC Class Codes: G06N 3/02 Neural networks
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06Q 20/203 Inventory monitoring
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0269 based on user profile or attribute
• CPC Class Codes: G06Q 30/0282 Rating or review of business operators or products
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34

Abstract

In selected embodiments a recommendation generator builds a network of interrelationships between venues, reviewers and users based on their attributes and reviewer and user reviews of the venues. Each interrelationship or link may be positive or negative and may accumulate with other links (or anti-links) to provide nodal links the strength of which are based on commonality of attributes among the linked nodes and/or common preferences that one node, such as a reviewer, expresses for other nodes, such as venues. The links may be first order (based on a direct relationship between, for instance, a reviewer and a venue) or higher order (based on, for instance, the fact that two venue are both liked by a given reviewer). The recommendation engine in certain embodiments determines recommended venues based on user attributes and venue preferences by aggregating the link matrices and determining the venues which are most strongly coupled to the user.

Claims

1. A system for generating recommendations for users from data networks, the system comprising:

processing circuitry; and

a non-transitory computer readable memory coupled to the processing circuitry, the memory storing machine-executable instructions, wherein the machine-executable instructions, when executed on the processing circuitry, cause the processing circuitry to

extract, from remote computing devices of a plurality of data sources,

a) first attribute data for a plurality of items, wherein the first attribute data comprises review data from a plurality of reviewers for a portion of the plurality of items,

b) second attribute data for the plurality of reviewers, and

c) third attribute data for a plurality of users requesting recommendations from the system,

wherein the extracted first, second, and third attribute data are stored on a non-transitory storage medium,

build, from the extracted first, second, and third attribute data, a data network comprising plurality of nodes and a plurality of links, each link of the plurality of links connecting two or more nodes of the plurality of nodes based on one or more interrelationships between the two or more nodes, wherein

the plurality of nodes comprise a plurality of item nodes each representing one of the plurality of items, a plurality of reviewer nodes each representing one of the plurality of reviewers, and a plurality of user nodes each representing one of the plurality of users, and

building the data network comprises

linking, within the non-transitory storage medium, each reviewer node of the plurality of reviewer nodes and each user node of the plurality of user nodes to at least one item node of the plurality of item nodes based on an affinity of the respective reviewer or respective user for at least one item of the plurality of items,

linking, within the non-transitory storage medium, each item node of two or more of the plurality of item nodes with a different item node of the plurality of item nodes to establish a respective set of linked item nodes based in part on a shared affinity of a reviewer of the plurality of reviewers or a user of the plurality of users for items represented by the set of linked item nodes, and

determining, for each set of linked item nodes, a respective link strength based in part on the shared affinity for the respective items by the respective reviewer or the respective user,

responsive to receiving a recommendation request from a remote computing device of a user of the plurality of users via a network, identify, using the data network, one or more item recommendations for the user corresponding to one or more recommended items of the plurality of items, wherein the one or more item recommendations are identified based in part on the respective link strengths connecting each item of the one or more recommended items within a respective set of linked item nodes, and

present, within a user interface screen at the remote computing device of the user responsive to receiving the recommendation request, the one or more item recommendations.

2. The system of claim 1, wherein the plurality of items comprise a plurality of venues.

3. The system of claim 2, wherein the plurality of venues comprise at least one of restaurants, hotels, or theaters.

4. The system of claim 1, wherein building the data network further comprises:

linking, within the non-transitory storage medium, each item node of two or more of the plurality of item nodes with a different item node of the plurality of item nodes to establish a respective set of linked item nodes based in part on shared attributes of respective items represented by the respective set of linked item nodes; and

determining, for each set of linked item nodes, a respective link strength based in part on the shared attributes of items represented by the respective set of linked item nodes.

5. The system of claim 1, wherein building the data network further comprises:

linking, within the non-transitory storage medium, each item node of two or more of the plurality of item nodes with a different item node of the plurality of item nodes to establish a respective set of linked item nodes based in part on shared attributes of the second attribute data of a respective two reviewers of the plurality of reviewers,

wherein

the respective two reviewers share affinities for items represented by the set of linked item nodes; and

determining, for each set of linked item nodes, a respective link strength based in part on the shared affinities for the respective items by the respective two reviewers.

6. The system of claim 1, wherein the extracted attribute data for each of the plurality of reviewers comprises at least one of gender, age, profession, or marital status.

7. The system of claim 1, wherein the respective link strength for each set of linked item nodes comprises a positive link strength when the reviewer or the user has substantially similar affinities for the items represented by the respective set of linked item nodes.

8. The system of claim 1, wherein the respective link strength for each set of linked item nodes comprises a negative link strength when the reviewer or the user has substantially different affinities for the items represented by the respective set of linked item nodes.

9. The system of claim 1, wherein the second attribute data for each of the plurality of reviewers and the third attribute data for each of the plurality of users comprises affinity data representing a preference of the respective reviewer or the respective user for one or more of the plurality of items.

10. The system of claim 9, wherein extracting the second attribute data and the third attribute data comprises

extracting, from remote computing devices for a portion of the plurality of data sources, financial transaction records for a portion of the plurality of reviewers and a portion of the plurality of users.

11. The system of claim 10, wherein the affinity data for a reviewer of the plurality of reviewers or a user of the plurality of users comprises a purchase history for one or more of the plurality items, wherein the purchase history is extracted from the financial transaction records.

12. The system of claim 9, wherein the machine-executable instructions, when executed on the processing circuitry, further cause the processing circuitry to:

determine, based in part on the affinity data, an amount of affinity of each reviewer of the plurality of reviewers for one or more respective items of the plurality of items.

13. The system of claim 12, wherein building the data network further comprises:

determining, for each set of linked item nodes, the respective link strength based in part on the amount of affinity of the respective reviewer for one or more respective items of the plurality of items.

14. A method comprising:

extracting, by processing circuitry from remote computing devices of a plurality of data sources,

a) first attribute data for a plurality of venues, wherein the first attribute data comprises review data from a plurality of reviewers for a portion of the plurality of venues,

b) second attribute data for the plurality of reviewers, and

c) third attribute data for a plurality of users requesting recommendations from the system;

storing, by the processing circuitry within a non-transitory storage medium, the extracted first, second, and third attribute data;

building, by the processing circuitry from the first, second, and third attribute data, a data network comprising a plurality of nodes and a plurality of links, each link of the plurality of links connecting two or more nodes of the plurality of nodes based on one or more interrelationships between the two or more nodes, wherein

the plurality of nodes comprises a plurality of venue nodes each representing one of the plurality of venues, a plurality of reviewer nodes each representing one of the plurality of reviewers, and a plurality of user nodes each representing one of the plurality of users, and

building the data network comprises

linking, within the non-transitory storage medium, each reviewer node of the plurality of reviewer nodes and each user node of the plurality of user nodes to at least one venue node of the plurality of venue nodes based on an affinity of the respective reviewer or respective user for at least one venue of the plurality of venues,

linking, within the non-transitory storage medium, each venue node of two or more of the plurality of venue nodes with a different venue node of the plurality of venue nodes to establish a respective set of linked venue nodes based in part on shared attributes of the second attribute data of a respective two reviewers of the plurality of reviewers, wherein

the respective two reviewers share affinities for venues represented by the set of venue nodes, and

determining, for each set of linked venue nodes, a respective link strength based in part on the shared affinities for the respective venues by the respective two reviewers;

responsive to receiving a recommendation request from a remote computing device of a user of the plurality of users via a network, identifying, by the processing circuitry using the data network, one or more venue recommendations for the user corresponding to one or more recommended venues of the plurality of venues, wherein the one or more venue recommendations are identified based in part on the respective link strengths connecting each venue of the one or more recommended venues within a respective set of linked venue nodes; and

presenting, by the processing circuitry within a user interface screen at the remote computing device of the user responsive to receiving the recommendation request, the one or more venue recommendations.

15. The method of claim 14, wherein building the data network further comprises:

linking, within the non-transitory storage medium, each venue node of two or more of the plurality of venue nodes with a different venue node of the plurality of venue nodes to establish a respective set of linked venue nodes based in part on a shared affinity of a reviewer of the plurality of reviewers or a user of the plurality of users for venues represented by the set of linked venue nodes, and

determining, for each set of linked venue nodes, a respective link strength based in part on the shared affinity for the respective venues by the respective reviewer or the respective user.

16. The method of claim 14, wherein building the data network further comprises:

linking, within the non-transitory storage medium, each venue node of two or more of the plurality of venue nodes with a different venue node of the plurality of venue nodes to establish a respective set of linked venue nodes based in part on shared attributes of respective venues represented by the respective set of linked venue nodes; and

determining, for each set of linked venue nodes, a respective link strength based in part on the shared attributes of venues represented by the two or more linked venue nodes.

17. The method of claim 14, wherein the second attribute data for each of the plurality of reviewers and the third attribute data for each of the plurality of users comprises affinity data representing a preference of the respective reviewer or the respective user for one or more of the plurality of venues.

18. The method of claim 17, wherein extracting the second attribute data and the third attribute data comprises extracting, from remote computing devices for a portion of the plurality of data sources, cellular positioning data for a portion of the plurality of reviewers and the plurality of users.

19. The method of claim 18, wherein the affinity data for a reviewer of the plurality of reviewers or a user of the plurality of users comprises a dwell time of the reviewer or the user at one or more of the plurality venues, wherein the dwell time is derived from the extracted cellular positioning data.

20. The method of claim 17, further comprising:

determining, by the processing circuitry based in part on the extracted affinity data, an amount of affinity of the respective reviewer for one or more venues of the plurality of venues; and

determining, by the processing circuitry for each set of linked venue nodes, the respective link strength based in part on the amount of affinity of the respective reviewer for at least one of the venues of the plurality of venues represented by the respective set of linked venue nodes.