HYBRID RECOMMENDATION SYSTEM

• US 20110010366A1
• Filed: 07/10/2009
• Published: 01/13/2011
• Est. Priority Date: 07/10/2009
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Tennenholtz, Moshe, Varshavsky, Roy, Karidi, Ron
• Granted Patent: US 8,661,050 B2
• Time in Patent Office: Days
• US Class Current: 707/732
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9536 Search customisation based on social or collaborative filtering
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

A recommendation system may use a network of relationships between many different entities to find search results and establish a relevance value for the search results. The relevance value may be calculated by analyzing trust and similarity components of each relationship between the search user and the entity providing the search results. The entities may be, for example, persons associated within express or implied social networks, or corporations or other organizations with a historical or other reputation. The relationships may be created through many different contact mechanisms and may be unidirectional, asymmetric bidirectional, or symmetric bidirectional relationships. The relationships may be different based on topic or other factors.

Claims

1. A method being performed on a computer processor, said method comprising:

receiving a search request from a search user;

performing an analysis of said search request to generate a category for said search request;

creating a network of entities, each of said entities having a relationship to at least one other entity, one of said entities representing said search user;

for each of said relationships, defining a trust component and a similarity component;

identifying a plurality of search results based on said search request, each of said plurality of search results being associated with a result entity, said result entity being one of said entities;

for each of said plurality of search results, determining a set of relationships between said search user and said result entities, and determining a relevance value based on said set of relationships;

ranking said plurality of search results based on said relevance value; and

presenting said plurality of search results according to said ranking.

2. The method of claim 1, said relationships being a function of said category.

3. The method of claim 2 further comprising:

said network being created prior to receiving said search request;

after receiving said search request, determining said relationships based on said category.

4. The method of claim 2 further comprising:

after receiving said search request, creating said network of entities.

5. The method of claim 2 further comprising:

determining a weighting for said trust component and said similarity component, said weighting being a function of said category.

6. The method of claim 1, said entities comprising persons in a social network.

7. The method of claim 1, at least one of said entities being an organization.

8. The method of claim 1, said network being limited to a predefined number of generations of relationships.

9. The method of claim 1 further comprising:

identifying a first entity;

determining a relationship between said first entity and a second entity, said relationship comprising a trust component and a similarity component; and

adding said entity and said relationship to a network database.

10. The method of claim 9, said entity being identified by membership in a social network.

11. The method of claim 9, said entity being identified by a communication between said first entity and said second entity.

12. The method of claim 1, said searching comprising a random walk search algorithm.

13. The method of claim 1, said creating a network being performed by a method comprising:

identifying a first set of entities having relationships with said one of said entities representing said search user;

adding said first set of entities to said network; and

for each of said entities in said first set of entities, identifying a set of entities having relationships to said each of said entities, and adding said set of entities to said network.

14. The method of claim 13, said creating a network being performed as a background process prior to said receiving said search request.

15. The method of claim 13, said creating a network being performed after said receiving said search request.

16. A system comprising:

a database comprising entities and relationships between entities;

for each of said relationships, defining a trust component and a similarity component;

a network creator configured to identify a starting entity and build a network based on said relationships;

a search manager configured to perform a method on a computer processor, said method comprising;

receiving a search request from a search user, said search user being a starting entity for said network;

performing an analysis of said search request to generate a category for said search request;

searching said network of entities to identify a plurality of search results based on said search request, each of said plurality of search results being associated with a result entity, said result entity being one of said entities;

for each of said plurality of search results, determining a set of relationships between said search user and said result entities, and determining a relevance value based on said set of relationships; and

ranking said plurality of search results based on said relevance value;

a user interface generator configured to presenting said plurality of search results according to said ranking on a user interface.

17. The system of claim 16 further comprising:

an entity manager configured to identify a new entity, determine a relationship between said new entity and another entity, and add said new entity to said database.

18. A method performed on a computer processor, said method comprising:

receiving a search request from a search user;

performing an analysis of said search request to generate a category for said search request;

determining a weighting scheme based on said category;

creating a network of entities, each of said entities having a relationship to at least one other entity, one of said entities representing said search user;

for each of said relationships, defining a trust component and a similarity component;

for each of said relationships, applying said weighting scheme to said trust component and said similarity component;

searching said network of entities to identify a plurality of search results based on said search request, each of said plurality of search results being associated with a result entity, said result entity being one of said entities;

for each of said plurality of search results, determining a set of relationships between said search user and said result entities, and determining a relevance value based on said set of relationships;

ranking said plurality of search results based on said relevance value; and

presenting said plurality of search results according to said ranking.

19. The method of claim 18 further comprising:

capturing a user selection, said user selection being one of said plurality of search results; and

adjusting at least one of said relationships based on said user selection to create an updated relationship.

20. The method of claim 19, said searching being performed by crawling said network.