BEHAVIOR TARGETING SOCIAL RECOMMENDATIONS

• US 20130304731A1
• Filed: 12/31/2010
• Published: 11/14/2013
• Est. Priority Date: 12/31/2010
• Status: Active Grant

• Current Assignee: Verizon Patent and Licensing Incorporated (Verizon Communications Inc.)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Zhao, Xianyu, Liu, Peng, Zheng, Hao
• Granted Patent: US 9,087,106 B2
• Time in Patent Office: Days
• US Class Current: 707/728
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: G06Q 30/0282 Rating or review of business operators or products
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

A process for generating social recommendations is provided. For each user, a user profile index is accessed to determine reading interests of the user. Further, relevance matching is performed to determine matching users having at least one publishing interest that is relevant to the reading interests of the user. Next, the matching users are ranked. Based on the ranking, one or more top ranked matching user(s) are determined. Additionally, a social recommendation for each of the top ranked matching user(s) is enabled to be made to the user.

Claims

1. A method, comprising:

for each user of a plurality of users;

accessing a user profile index to determine at least one reading interest of the user;

indexing publishing interests of the user based on gathered publishing data from the user, such that the publishing interests include topics generated from key words in the publishing behavior information and latent semantic topics inferred from the publishing behavior information;

performing relevance matching to determine matching users such that each matching user has at least one publishing interest that is relevant to at least one reading interest of the user;

ranking the matching users;

based on the ranking, determining one or more top ranked matching users; and

enabling a social recommendation for each of the one or more top ranked matching users to be made to the user.

2. The method of claim 1, further comprising:

making a suggestion to the user to friend at least one of the one or more top ranked matching users.

3. The method of claim 1, further comprising:

making a suggestion to the user to follow at least one of the one or more top ranked matching users.

4. The method of claim 1, further comprising:

gathering consumption behavior information from the plurality of users; and

gathering publishing behavior information from the plurality of users.

5. The method of claim 4, wherein the consumption behavior information includes at least one of page views, ad views, ad clicks, or search queries.

6. The method of claim 5, wherein the publishing behavior information includes at least one of blogs, tweets, articles, or published user-generated content.

7. The method of claim 1, further comprising:

generating the user profile index by indexing a plurality of users profiles based on gathered consumption behavior from the plurality of users such that the user profile index includes reading interests of the users;

indexing publishing interests of the plurality of users based on gathered publishing data from the plurality of users.

8. The method of claim 7, wherein the indexing of the publishing interests is accomplished such that the publishing interests include topics generated from key words in the publishing behavior information and latent semantic topics inferred from the publishing behavior information by a probabilistic model using latent Dirichlet allocation, and wherein the indexing of the plurality of user profiles is accomplished such that the reading interests employ a weighted aggregation of topic distribution vectors such that the weighting is set in an exponential time-decaying manner.

9. A network device, comprising:

a transceiver that is configured to send and receive data over a network; and

a processor that is configured to perform actions, including

for each user of a plurality of users;

accessing a user profile index to determine at least one reading interest of the user;

indexing publishing interests of the user based on gathered publishing data from the user, such that the publishing interests include topics generated from key words in the publishing behavior information and latent semantic topics inferred from the publishing behavior information;

performing relevance matching to determine matching users such that each matching user has at least one publishing interest that is relevant at least one reading interest of the user;

ranking the matching users;

based on the ranking, determining one or more top ranked matching users; and

enabling a social recommendation for each of the one or more top ranked matching users to be made to the user.

10. The network device of claim 9, wherein

ranking the matching users is done based on at least one of:

freshness of published content of the matching user, popularity of the published content of the matching user, relevance of the published content of the matching user, expertise of the matching user, demographic information of the user, or at least one descriptive statistic of the user.

11. The network device of claim 9, wherein the relevance matching is accomplished with an information retrieval process and a look-up table.

12. A tangible processor-readable medium having processor-executable code encoded therein, which when executed by one or more processors, enables actions, comprising:

for each user of a plurality of users

accessing a user profile index to determine at least one reading interest of the user;

indexing publishing interests of the user based on gathered publishing data from the user, such that the publishing interests include topics generated from key words in the publishing behavior information and latent semantic topics inferred from the publishing behavior information;

performing relevance matching to determine matching users such that each matching user has at least one publishing interest that is relevant at least one reading interest of the user;

ranking the matching users;

based on the ranking, determining one or more top ranked matching users; and

enabling a social recommendation for each of the one or more top ranked matching users to be made to the user.

13. The tangible processor-readable medium of claim 12, wherein

ranking the matching users is done based on at least one of:

freshness of published content of the matching user, popularity of the published content of the matching user, relevance of the published content of the matching user, expertise of the matching user, demographic information of the user, or at least one descriptive statistic of the user.

14. The tangible processor-readable medium of claim 12, wherein the relevance matching is accomplished with an information retrieval process and a look-up table.

15. A system, comprising:

a plurality of network devices that are configured to perform actions, including;

for each user of a plurality of users;

accessing a user profile index to determine at least one reading interest of the user;

indexing publishing interests of the user based on gathered publishing data from the user, such that the publishing interests include topics generated from key words in the publishing behavior information and latent semantic topics inferred from the publishing behavior information;

performing relevance matching to determine matching users such that each matching user has at least one publishing interest that is relevant to at least one reading interest of the user;

ranking the matching users;

based on the ranking, determining one or more top ranked matching users; and

enabling a social recommendation for each of the one or more top ranked matching users to be made to the user.

16. The system of claim 15, wherein the plurality of network devices are configured to perform further actions, including:

making a suggestion to the user to friend at least one of the one or more top ranked matching users.

17. The system of claim 15, wherein the plurality of network devices are configured to perform further actions, including:

making a suggestion to the user to follow at least one of the one or more top ranked matching users.

18. The system of claim 15, wherein the plurality of network devices are configured to perform further actions, including:

generating the user profile index by indexing a plurality of users profiles based on gathered consumption behavior from the plurality of users such that the user profile index includes reading interests of the users;

indexing publishing interests of the plurality of users based on gathered publishing data from the plurality of users.

19. The system of claim 15, wherein the plurality of network devices are configured to perform further actions, including:

gathering consumption behavior information from the plurality of users; and

gathering publishing behavior information from the plurality of users.

20. The system of claim 19, wherein the plurality of network devices are further configured such that the consumption behavior information includes at least one of page views, ad views, ad clicks, or search queries, and wherein the publishing behavior information includes at least one of blogs, tweets, articles, or published user-generated content.