Social network recommended content and recommending members for personalized search results

• US 8,949,232 B2
• Filed: 10/04/2011
• Issued: 02/03/2015
• Est. Priority Date: 10/04/2011
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Harrington, Timothy, Shenoy, Rajesh, Najork, Marc, Panigrahy, Rina
• Primary Examiner(s): TRUONG, CAM Y T
• Application Number:
  US13/252,215
• Publication Number:
  US 20130086057A1
• Time in Patent Office: 1,218 Days
• Field of Search: 707/802, 707/732
• US Class Current: 707/732
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: H04L 51/52 for supporting social networking services

Abstract

Architecture that provides a data structure to facilitate personalized ranking over recommended content (e.g., documents). The data structure approximates the social distance of the searching user to the content at query time. A graph is created of content recommended by members of the social network, where the nodes of the graph include content nodes (for the content) and recommending member nodes (for members of the social network who recommended the content). If a member recommends content, an edge is created between the member node and the content node. If a member is a “friend” (tagged as related in some way) of another member, an edge is created between the two member nodes. Each node is converted to a lower dimensional feature set. Feature sets of the content are indexed and the feature set of the searching user is utilized to match and rank the search results at query time.

Claims

1. A method, comprising:

creating a graph of nodes and node relationships, the graph comprising content nodes of content recommended by members of a social network, entity nodes of the recommending members, and links between the entity nodes according to social links between associated members in the social network, each content node being configured so that the content node (i) does not link to another content node and (ii) links to only a single entity node, the single entity node being the entity node of the recommending member who has recommended the content of the content node;

converting each node to a feature set of auxiliary information;

sampling some sets of the nodes, the nodes of the sampled sets being seed nodes, each feature set including (i) information about seed nodes closest to the node from which the feature set was converted and (ii) distances to those seed nodes, a distance between a pair of feature sets approximating a social distance between the nodes from which the feature sets were converted;

indexing the node relationships for keyword searches that return recommended content of the content nodes;

identifying a searching user as associated with the social network; and

processing a query against the index using keywords.

2. The method of claim 1, further comprising creating a content identifier for each content node and an entity identifier for each entity node.

3. The method of claim 2, further comprising indexing the content identifier, and storing associated content and associated feature set as searched keywords.

4. The method of claim 3, further comprising converting the feature set into searchable keywords.

5. The method of claim 2, further comprising storing the entity identifier in a key-value store, and searching the key-value store for a feature set based on the entity identifier.

6. The method of claim 1, further comprising looking-up a feature set for the searching user based on a corresponding entity identifier.

7. The method of claim 1, further comprising combining the query with a social circle query to find search results that include results from the social network.

8. The method of claim 1, further comprising approximating social distance between the entities associated with the searching user and the recommended content.

9. The method of claim 8, wherein the social distance is based on attributes of the user.

10. The method of claim 9, wherein content recommended by another user is ranked higher when the another user is at the same location as the user.

11. The method of claim 1, wherein each feature set is a vector of the auxiliary information.

12. A method, comprising:

creating a graph of content nodes, entity nodes, and relationships between the nodes, (i) the content nodes being nodes of content recommended by members of a social network, each content node being linkable to only a single entity node, which is the entity node of a member who recommends the content of that content node, and (ii) the entity nodes being of the recommending members, each entity node being linkable to both content nodes and other entity nodes, links between the entity nodes being based on social links between associated members in the social network, converting each node to a feature set of auxiliary information;

sampling some sets of the nodes, the nodes of the sampled sets being seed nodes, each feature set including (i) information about seed nodes closest to the node from which the feature set was converted and (ii) distances to those seed nodes, a distance between a pair of feature sets approximating a social distance between the nodes from which the feature sets were converted;

indexing the node relationships of the created graph for keyword searches that return recommended content of the content nodes;

identifying a searching user as associated with the social network; and

processing a query against the index using keywords.

13. A method, comprising:

creating a graph of nodes and node relationships, the graph consisting of content nodes of content recommended by members of a social network, entity nodes of the recommending members, and links between the entity nodes according to social links between associated members in the social network, each content node being configured so that the content node is linkable only to a single entity node, the single entity node being the entity node of the recommending member who has recommended the content of the content node;

converting each node to a feature set of auxiliary information;

sampling some sets of the nodes, the nodes of the sampled sets being seed nodes, each feature set including (i) information about seed nodes closest to the node from which the feature set was converted and (ii) distances to those seed nodes, a distance between a pair of feature sets approximating a social distance between the nodes from which the feature sets were converted;

indexing the node relationships of the graph for keyword searches that return recommended content of the content nodes to yield an index; and

processing a query submitted by a user who is associated with the social network against the index using keywords.