Recommending content items in a social network using delayed interaction

• US 10,430,421 B2
• Filed: 12/29/2014
• Issued: 10/01/2019
• Est. Priority Date: 12/29/2014
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Inventors: Keggi, John Graham, Giambalvo, Daniel Joseph, Xue, Mao
• Primary Examiner(s): Burke, Jeff A
• Assistant Examiner(s): Dewan, Kamal K
• Application Number:
  US14/585,085
• Publication Number:
  US 20160188596A1
• Time in Patent Office: 1,737 Days
• Field of Search: 707758, 707705, 707723, 707728, 707732, 707737, 707755, 715206, 715826
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 30/0631 Item recommendations
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

When a user sees a content item presented by a social networking system, the user may select an option to save the content item. When a user saves a content item, views saved content items, or otherwise indicates a present interest in a particular saved content item, the system recommends one or more additional items for the users to consume or save based on the seed saved content item. To find the additional content items, the system identifies other users who also saved the seed item and then finds other content items that these other users also saved at a rate that is disproportionately higher than the global save rates for the content items (which may be normalized by opportunities to save the content). Relevance for content items in other contexts may also be determined based on content items that have been saved by a user.

Claims

1. A computer-implemented method comprising:

providing content items for display to users of a social networking system, wherein at least some of the content items are displayed with an input option providing the users an opportunity to save the content item for subsequent presentation to the users that saved the content item;

receiving a request to save content items from users presented with the opportunity to save the content items;

receiving a request for content items related to a seed content item in the social networking system;

identifying a set of users that saved the seed content item;

identifying a set of co-saved content items that are also saved by the set of users that saved the seed content item;

for each co-saved content item of the set of co-saved content items, calculating a relevance score for the co-saved content item with respect to the seed content item by;

for a first population of users that were provided with an opportunity to save the co-saved content item and were not provided with an opportunity to save the seed content item, identifying a first percentage of the first population of users that saved the co-saved content item,

for a second population of users that saved the seed content item and were provided with an opportunity to save both the co-saved content item and the seed content item, identifying a second percentage of the second population of users that saved the co-saved content item, and

calculating the relevance score for the co-saved content item based on a comparison of the first percentage of the first population of users with the second percentage of the second population of users;

ranking the co-saved content items based on the relevance scores of the co-saved content items;

selecting one or more of the co-saved content items based on the ranking; and

providing the selected one or more co-saved content items for display to a user.

2. The method of claim 1, wherein the received request for content items is associated with a user of the social networking system, and the method further comprises filtering the set of co-saved content items to exclude content items that are not accessible to the user based on privacy settings of the content items.

3. The method of claim 1, further comprising filtering the set of co-saved content items to exclude content items that do not match a content type of the seed content item.

4. The method of claim 3, wherein the content type comprises events, articles, or videos.

5. The method of claim 1, further comprising filtering the set of co-saved content items to exclude content items stored by the social networking system.

6. The method of claim 1, further comprising filtering the set of co-saved content items to exclude content items stored at an external system to the social networking system.

7. The method of claim 1, wherein the request for content items is received responsive to a user saving the seed content item.

8. The method of claim 1, wherein the request for content items is received responsive to a user interacting with the seed content item.

9. The method of claim 1, wherein the request for content items is received responsive to a user selecting the seed content item in a list of saved content items, and the method further comprises filtering the set of co-saved content items to exclude content items that are not in the list of saved content items.

10. A method comprising:

providing content items for display to a user of a social networking system, wherein at least some of the content items are displayed with an input option providing the user an opportunity to save the content item for later consumption;

receiving a request to save content items from the user presented with the opportunity to save the content items;

receiving a request to determine the relevance of a seed content item to the user; and

generating a relevance score based on the similarity of the seed content item to the content items saved by the user.

11. The method of claim 10, further comprising receiving a context of the user, the context describing at least one of a time, a location of the user, and a device of the user;

and wherein generating the relevance score is further based on the context of the user.

12. The method of claim 11, further comprising generating a behavioral model of user behaviors based on user consumption of the saved content items, and wherein generating the relevance score based on the context of the user is based on the behavioral model.

13. The method of claim 12, wherein the behavioral model is a trained computer model.

14. The method of claim 10, further comprising identifying one or more content items of the saved content items that were consumed by the user;

wherein generating the relevance score comprises increasing a weighting of the one or more consumed content items.

15. A non-transitory computer-readable medium comprising instructions for execution on a processor, the instructions causing the processor to perform steps of:

providing content items for display to users of a social networking system, wherein at least some of the content items are displayed with an input option providing the users an opportunity to save the content item for subsequent presentation to the users that saved the content item;

receiving a request to save content items from users presented with the opportunity to save the content items;

receiving a request for content items related to a seed content item in the social networking system;

identifying a set of users that saved the seed content item;

identifying a set of co-saved content items that are also saved by the set of users that saved the seed content item;

for each co-saved content item of the set of co-saved content items, calculating a relevance score for the co-saved content item with respect to the seed content item by;

for a first population of users that were provided with an opportunity to save the co-saved content item and were not provided with an opportunity to save the seed content item, identifying a first percentage of the first population of users that saved the co-saved content item,

for a second population of users that saved the seed content item and were provided with an opportunity to save both the co-saved content item and the seed content item, identifying a second percentage of the second population of users that saved the co-saved content item, and

calculating the relevance score for the co-saved content item based on a comparison of the first percentage of the first population of users with the second percentage of the second population of users;

ranking the co-saved content items based on the relevance scores of the co-saved content items;

selecting one or more of the co-saved content items based on the ranking; and

providing the selected one or more co-saved content items for display to a user.

16. The non-transitory computer-readable medium of claim 15, wherein the received request for content items is associated with a user of the social networking system, and the method further comprises filtering the set of co-saved content items to exclude content items that are not accessible to the user based on privacy settings of the content items.

17. The non-transitory computer-readable medium of claim 15, the steps further including filtering the set of co-saved content items to exclude content items that do not match a content type of the seed content item.

18. The non-transitory computer-readable medium of claim 17, wherein the content type comprises events, articles, or videos.

19. The non-transitory computer-readable medium of claim 15, the steps further comprising filtering the set of co-saved content items to exclude content items stored by the social networking system.

20. The non-transitory computer-readable medium of claim 15, the steps further comprising filtering the set of co-saved content items to exclude content items stored at an external system to the social networking system.

21. The non-transitory computer-readable medium of claim 15, wherein the request for content items is received responsive to a user saving the seed content item.

22. The non-transitory computer-readable medium of claim 15, wherein the request for content items is received responsive to a user interacting with the seed content item.

23. The non-transitory computer-readable medium of claim 15, wherein the request for content items is received responsive to a user selecting the seed content item in a list of saved content items, and the method further comprises filtering the set of co-saved content items to exclude content items that are not in the list of saved content items.