PROBABILISTIC RECOMMENDATION OF AN ITEM

• US 20110184806A1
• Filed: 01/27/2010
• Published: 07/28/2011
• Est. Priority Date: 01/27/2010
• Status: Active Grant

• Current Assignee: eBay Inc.
• Original Assignee: eBay Inc.
• Inventors: Chen, Ye, Canny, John
• Granted Patent: US 9,760,802 B2
• Time in Patent Office: Days
• US Class Current: 705/14.52
• CPC Class Codes: G06F 18/2321 using statistics or function optimisation, e.g. modelling of probability density functions
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0251 Targeted advertisements
• CPC Class Codes: G06Q 30/0254 based on statistics
• CPC Class Codes: G06Q 30/0257 User requested
• 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

Abstract

A clustering and recommendation machine determines that an item is included in a cluster of items. The machine accesses item data descriptive of the item. The machine accesses a vector that represents the cluster and calculates the likelihood that the item is included in the cluster, based on the item variable and the probability parameter. The machine determines that the item is included in the cluster, based on the likelihood. The machine also recommends an item to a potential buyer. The machine accesses behavior data that represents a first event type pertinent to a first cluster of items. The machine calculates a probability that a second event type pertaining to a second cluster of items will co-occur with the first event type. The machine identifies an item from the second cluster to be recommended and presents a recommendation of the item to the potential buyer.

Claims

1. A computer-implemented method of recommending an item to a potential buyer, the method comprising:

accessing behavior data pertinent to a first cluster of items, the first cluster representing a first product of which each item of the first cluster is a specimen, the behavior data including an event record representative of a first event type pertinent to the first product;

calculating a probability based on the behavior data, the probability being of a co-occurrence of a second event type with the first event type, the second event type pertaining to a second product represented by a second cluster of items, the second cluster including the item to be recommended to the potential buyer, the calculating being performed by a module implemented using a processor of a machine;

identifying the item based on the probability of the co-occurrence; and

presenting a recommendation of the item to the potential buyer, the recommendation including item data descriptive of the item and indicating the item as a specimen of the second product.

2. The computer-implemented method of claim 1, wherein:

the first event type is a purchase of a specimen of the first product; and

the second event type is a purchase of a specimen of the second product.

3. The computer-implemented method of claim 1, wherein:

the first event type is a bid on a specimen of the first product; and

the second event type is a bid on a specimen of the second product.

4. The computer-implemented method of claim 1, wherein:

the first event type is a click to request information pertinent to the first product; and

the second event type is a bid on a specimen of the second product.

5. The computer-implemented method of claim 1, wherein:

the first event type is a page view event referencing information pertinent to the first product; and

the second event type is a bid on a specimen of the second product.

6. The computer-implemented method of claim 1 further comprising:

accessing popularity data pertinent to the first product; and

wherein

the identifying of the item is based on the popularity data.

7. The computer-implemented method of claim 1, wherein:

the calculating of the probability includes generating a matrix based on the behavior data, the matrix representing a pattern of co-occurrence pertinent to the first and second event types.

8. The computer-implemented method of claim 1 further comprising:

determining that an instance of the first event type and an instance of the second event type occurred within a threshold time period.

9. The computer-implemented method of claim 1, wherein:

the behavior data includes position data of a hyperlink presented in a web page, the hyperlink being to request information pertinent to the first product; and

the calculating of the probability is based on the position data.

10. The computer-implemented method of claim 1 further comprising

removing a portion of the behavior data generated by a user, the removing being based on a number of activities performed by the user within a threshold time period.

11. The computer-implemented method of claim 1 further comprising

determining a rank of the item to be recommended to the potential buyer;

and wherein

the presenting of the recommendation is based on the rank of the item.

12. The computer-implemented method of claim 11 further comprising

accessing a deadline of the item, the deadline indicating a time after which the item is unavailable for purchase;

and wherein

the determining of the rank is based on the deadline of the item.

13. The computer-implemented method of claim 1 further comprising:

calculating an argument of a maximum of the probability; and

wherein

the identifying of the item is based on the argument of the maximum.

14. The computer-implemented method of claim 1 further comprising:

accessing a trust score of a seller of the item; and

wherein

the identifying of the item is based on the trust score of the seller.

15. A system to recommend an item to a potential buyer, the system comprising:

an access module to access behavior data pertinent to a first cluster of items, the first cluster representing a first product of which each item of the first cluster is a specimen, the behavior data including an event record representative of a first event type pertinent to the first product;

a hardware-implemented probability module to calculate a probability based on the behavior data, the probability being of a co-occurrence of a second event type with the first event type, the second event pertaining to a second product represented by a second cluster of items, the second cluster including the item to be recommended to the potential buyer;

a recommendation module to;

identify the item based on the probability of the co-occurrence; and

present a recommendation of the item to the potential buyer, the recommendation including item data descriptive of the item and indicating the item as a specimen of the second product.

16. The system of claim 15, wherein:

the first event type is selected from a first group consisting of;

a purchase of a specimen of the first product;

a bid on a specimen of the first product;

a click to request information pertinent to the first product; and

a page view event referencing information pertinent to the first product; and

the second event type is selected from a second group consisting of;

a purchase of a specimen of the second product; and

a bid on a specimen of the second product.

17. The system of claim 15, wherein

the probability module, in calculating the probability, is to generate a matrix based on the behavior data, the matrix representing a pattern of co-occurrence pertinent to the first and second event types

18. A machine-readable storage medium comprising instructions that, when executed by one or more processors of a machine, cause the machine to perform a method comprising:

accessing behavior data pertinent to a first cluster of items, the first cluster representing a first product of which each item of the first cluster is a specimen, the behavior data including an event record representative of a first event type pertinent to the first product;

calculating a probability based on the behavior data, the probability being of a co-occurrence of a second event type with the first event, the second event type pertaining to a second product represented by a second cluster of items, the second cluster including an item to be recommended to a potential buyer;

identifying the item based on the probability of the co-occurrence; and

presenting a recommendation of the item to the potential buyer, the recommendation including item data descriptive of the item and indicating the item as a specimen of the second product.

19. The machine-readable storage medium of claim 18, wherein the instructions, when executed by the one or more processors of the machine, cause the machine to perform the method, wherein:

the behavior data includes position data of a hyperlink presented in a web page, the hyperlink being to request information pertinent to the first product; and

the calculating of the probability is based on the position data.

20. The machine-readable storage medium of claim 18, wherein the instructions, when executed by the one or more processors of the machine, cause the machine to perform the method further comprising:

calculating an argument of a maximum of the probability; and

wherein

the identifying of the item is based on the argument of the maximum.