Systems and methods for trend aware self-correcting entity relationship extraction

• US 8,838,521 B2
• Filed: 08/16/2013
• Issued: 09/16/2014
• Est. Priority Date: 03/23/2010
• Status: Active Grant

• Current Assignee: eBay Inc.
• Original Assignee: eBay Inc.
• Inventors: Sinha, Ayon
• Primary Examiner(s): CHEN, ALAN S
• Application Number:
  US13/969,211
• Publication Number:
  US 20130339289A1
• Time in Patent Office: 396 Days
• US Class Current: 706/52
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: G06Q 30/0631 Item recommendations
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

Methods and systems for trend aware self-correcting entity relationship extraction are disclosed. For example, a method can include receiving a selected entity, determining a plurality of entities related to the selected entity, determining a plurality of most probable entities, calculating relevance scores, and displaying a subset of the plurality of most probable entities. The selected entity can be received on a network-based transaction system. The plurality of entities related to the selected entity can be determined based on a relationship score. The relationship score can represent navigation transitions, aggregated over time, between the selected entity and each of the plurality of entities. The plurality of most probable entities can be determined based on probabilities. Relevance scores can be calculated for each of the plurality of most probable entities. Finally, the subset of the plurality of most probable entities to be displayed can be determined according to the relevance scores.

Claims

1. A method comprising:

receiving a selected entity on a network-based transaction system;

determining, using one or more processors on the network-based transaction system, a plurality of related entities based at least in part on edge strengths between the selected entity and a plurality of entities listed on the network-based transaction system;

selecting, from the plurality of related entities, a sub-set of related entities based at least in part on probabilities calculated from the edge strengths between the selected entity and the plurality of related entities; and

generating a recommendation set of related entities, for display via the network-based transaction system, from the sub-set of related entities based at least in part on a velocity for each related entity of the sub-set of related entities, wherein the velocity is a measure of directional change in popularity.

2. The method of claim 1, wherein generating the recommendation set includes calculating a relevance score for each related entity in the sub-set of related entities, wherein the relevance score is a function of probability and velocity.

3. The method of claim 2, wherein calculating the relevance score includes determining a category associated with the selected entity and filtering the sub-set of related entities in accordance with the category.

4. The method of claim 1, wherein determining the plurality of related entities includes calculating the edge strengths between the selected entity and the plurality of entities based on community behavior tracked by the network-based transaction system.

5. The method of claim 4, wherein the community behavior tracked by the network-based transaction system includes navigation between entities, and wherein calculating the edge strengths includes determining an amount of navigation between entities.

6. The method of claim 5, wherein determining the amount of navigation between entities can include determining user session boundaries for determining related navigation history.

7. The method of claim 6, wherein determining user session boundaries includes imposing artificial session boundaries based on at least one of idle time or browsing behavior.

8. A system comprising:

an activity tracking module configured to receive a selected entity on a network-based transaction system;

a relationship module configured to,

determine a plurality of related entities based at least in part on edge strengths between the selected entity and a plurality of entities listed on the network-based transaction system; and

select, from the plurality of related entities, a sub-set of related entities based at least in part on probabilities calculated from edge strengths between the selected entity and the plurality of related entities and based at least in part using an entity velocity for each of the plurality of related entities, wherein the entity velocity is a measure of a change over time in popularity; and

a recommendation engine configured to generate a recommendation set of related entities, for display on the network-based transaction system, from the sub-set of related entities.

9. The system of claim 8, wherein the relationship module is configured to determine the edge strengths between the selected entity and the plurality of entities at least in part by aggregating navigation transitions between entities.

10. The system of claim 9, wherein the relationship module is configured to weight more recent navigation transitions more heavily than older navigation transitions when aggregating the navigation transitions between entities.

11. The system of claim 8, wherein each probability of the probabilities calculated from edge strengths is calculated by normalizing navigation transitions between the selected entity and each entity of the plurality of entities using a plurality of navigation transitions associated with the selected entity.

12. The system of claim 8, wherein the relationship module is configured to calculate the change over time in popularity of the entity using a change in sales volume associated with the entity between a current time period and a past time period.

13. The system of claim 8, wherein the relationship module is configured to calculate the change over time in popularity of the entity using a change in page views associated with the entity between a current time period and a past time period.

14. A non-transitory machine-readable storage medium containing instructions that, when executed on a machine, cause the machine to perform operations comprising:

determine a plurality of related entities based at least in part on edge strengths between a selected entity and a plurality of entities listed on a network-based transaction system;

select, from the plurality of related entities, a sub-set of related entities based at least in part on probabilities calculated from edge strengths between the selected entity and the plurality of related entities; and

generate a recommendation set of related entities, for display on the network-based transaction system, from the sub-set of related entities based at least in part on a velocity for each related entity of the sub-set of related entities, wherein the velocity is a measure of directional change in popularity.

15. The non-transitory machine-readable storage medium of claim 14, wherein the instructions that cause the machine to generate the recommendation set include instructions that cause the machine to calculate a relevance score for each related entity in the sub-set of related entities, wherein the relevance score is a function of probability and velocity.

16. The non-transitory machine-readable storage medium of claim 15, wherein the instructions that cause the machine to calculate the relevance score include instructions that cause the machine to determine a category associated with the selected entity and filtering the sub-set of related entities in accordance with the category.

17. The non-transitory machine-readable storage medium of claim 14, wherein the instructions that cause the machine to determine the plurality of related entities include instructions that cause the machine to calculate the edge strengths between the selected entity and the plurality of entities based on community behavior tracked by the network-based transaction system.

18. The non-transitory machine-readable storage medium of claim 17, wherein the community behavior tracked by the network-based transaction system includes navigation between entities, and wherein calculating the edge strengths includes determining an amount of the navigation between entities and timing of the navigation between entities.