System for facilitating discovery and management of feeds

• US 7,984,056 B1
• Filed: 12/28/2007
• Issued: 07/19/2011
• Est. Priority Date: 12/28/2007
• Status: Expired due to Fees

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Kane, Francis J.
• Primary Examiner(s): Vital; Pierre M
• Assistant Examiner(s): OBISESAN, AUGUSTINE KUNLE
• Application Number:
  US11/966,816
• Time in Patent Office: 1,299 Days
• Field of Search: None
• US Class Current: 707/749
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages
• CPC Class Codes: H04L 51/52 for supporting social networking services

Abstract

A web feed portal can enable users to discover new web feeds and manage existing subscriptions of feeds. Recommended feeds can be generated for a user that reflect the user'"'"'s interests or that are popular within a user community. Users can subscribe to feeds by selecting recommended web feeds or by contributing web feeds to the web feed portal. The user'"'"'s web feed subscriptions can also be ranked to display the feeds in order of update time and/or time viewed by the user.

Claims

1. A system for recommending web feeds to subscribe to, the system comprising:

a computer system comprising one or more processors, the computer system programmed to implement;

a user database comprising user activity data regarding one or more web feed subscriptions of a target user, each web feed subscription of the target user corresponding to a web feed having a first address stored in a web feed repository;

a feed ranking module configured to generate a ranked set of at least a portion of the target user'"'"'s web feed subscriptions, wherein rankings in the ranked set are based at least in part on one or both of the duration since the feed was last updated and the duration since the user last viewed the feed;

an associations mining module configured to detect behavioral associations between first web feeds based at least partly on behavior of a plurality of users, wherein the associations mining module is configured to perform said detection by at least identifying pairs of the first web feeds that were both selected for subscription or viewing by a plurality of users, the association mining module additionally configured to create item-to-item mappings between the identified first web feeds, the item-to-item mappings comprising scores reflecting the degree of the associations;

a recommendations service configured to generate personalized web feed recommendations for the target user by at least accessing the behavioral associations to identify one or more of the identified first web feeds that are associated with a subset of the user activity data of the target user, and selecting a most highly-scored subset of the associated web feeds as the web feed recommendations; and

a page generator configured to generate a web feed user interface configured to display the ranked set of at least a portion of the target user'"'"'s web feed subscriptions and the personalized web feed recommendations, wherein selection of a recommended web feed by the target user subscribes the target user to the recommended web feed,

wherein the web feed user interface comprises a text field providing functionality for the target user to input text corresponding to a second address of a second web feed to thereby subscribe the target user to the second web feed and make the second web feed available for other users to select for subscription, and wherein the second address of the second web feed is added to the web feed repository if the second address is not already stored in the web feed repository;

the associations mining module being further configured to detect one or more additional behavioral associations between at least the second web feed and one or more of the first web feeds; and

the recommendations service being further configured to use the one or more additional behavioral associations to recommend the second web feed to one or more of the other users.

2. The system of claim 1, wherein the associations mining module configured to determine the scores at least partly based on the number of users that selected both of the associated web feeds for subscription or viewing.

3. The system of claim 1, wherein item-to-item mappings between web feeds are stored in an associated items table in computer storage.

4. The system of claim 1, wherein the recommendations service is further configured to generate the personalized web feed recommendations using a web feed represented by the web feed address submitted by the target user.

5. A computer-implemented method of recommending feeds to users, the method comprising:

by a computer system comprising one or more processors;

providing a web feed user interface comprising a text field providing functionality for a target user to input text corresponding to an address of a first web feed, the web feed user interface further providing functionality for the target user to submit a first web feed address to a web feed repository when the first web feed address is not already stored in the web feed repository, wherein inputting text corresponding to the address of the first web feed by the target user makes the web feed address available for other users to select for subscription;

detecting one or more behavioral associations between the first web feed and one or more second web feeds based at least partly on behavior of a first plurality of users, wherein said detecting comprises creating first item-to-item mappings between the first web feed and the one or more second web feeds in response to determining that the first web feed and the one or more second web feeds were both selected for subscription or viewing by the first plurality of users, the first item-to-item mappings comprising first scores reflecting a degree of association between the first web feed and the one or more second web feeds;

recommending the first web feed to one or more other users based at least partly on the one or more behavioral associations;

determining whether behavioral data exists for the target user, the behavioral data selected from at least one of the following;

(a) feeds that the target user has subscribed to and (b) feeds that the target user has viewed;

when behavioral data exists for the target user, generating a set of personalized web feed recommendations comprising a set of web feeds for the target user based at least in part on the behavioral data related to the user, wherein generating the personalized web feed recommendations comprises;

identifying pairs of web feeds stored in the web feed repository that were both selected for subscription or viewing by a second plurality of users and creating second item-to-item mappings between the identified web feeds, the item-to-item mappings comprising second scores reflecting a degree of associations between the identified web feeds,

identifying associated web feeds that are associated with a subset of the behavioral data of the target user, and

selecting a most highly-scored subset of the associated web feeds as the web feed recommendations; and

when no behavioral data exists for the target user, generating other feed recommendations for the target user based on criteria other than behavioral data.

6. The method of claim 5, wherein generating other feed recommendations for the target user comprises randomly selecting feeds to recommend to the target user.

7. The method of claim 5, wherein the other feed recommendations are generated based at least in part on feed popularity.

8. The method of claim 5, wherein an act of the user contributing one or more feeds to the feed database subscribes the user to the one or more feeds.

9. The method of claim 5, wherein providing a feed user interface comprises providing a listing of feed recommendations for the user, wherein selecting a selected feed of the feed recommendations in the listing subscribes the user to the selected feed.

10. The system of claim 5, further comprising adding subscribed feeds to a feed repository if the subscribed feeds are not already stored in the feed repository.

11. A computer-implemented method of generating web feed recommendations, the method comprising:

by a computer system comprising one or more processors;

receiving a request for web feed recommendations, the request including an identifier of a user; and

in response to the request,

retrieving behavioral data of the user;

selecting a set of web feeds to recommend to the user, wherein the web feeds are selected by at least;

identifying web feeds that were selected for subscription or viewing by a first plurality of users and creating first item-to-item mappings between the identified web feeds, the first item-to-item mappings comprising first scores reflecting the degree of the associations, wherein the first scores are determined at least partly based on the number of users that selected both of the associated identified web feeds for subscription or viewing, and

selecting a subset of the identified web feeds based at least partly on the behavioral data of the user and the scores of the identified web feeds;

generating a display page for presentation to the user, the display page including a text field providing functionality for the user to input a web feed address corresponding to a second web feed;

providing functionality to submit a web feed address to a web feed repository, wherein inputting a web feed address by the user makes the second web feed available for other users to select for subscription if the web feed address is not already stored in the web feed repository;

returning the personalized web feed recommendations;

detecting one or more behavioral associations between the second web feed and one or more web feeds stored in the web feed repository based at least partly on behavior of a second plurality of users, wherein said detecting comprises creating second item-to-item mappings between the second web feed and the one or more stored web feeds when the second web feed and the one or more stored web feeds were both selected for subscription or viewing by the second plurality of users, the second item-to-item mappings comprising second scores reflecting a degree of association between the second web feed and the one or more stored web feeds;

recommending the second web feed to one or more other users based at least partly on the one or more behavioral associations.

12. The method of claim 11, wherein selecting a set of web feeds to recommend to the user comprises selecting web feeds randomly to recommend to the user.

13. The method of claim 11, wherein receiving the request from the remote website comprises receiving a call to one or more routines in an application programming interface (API).

14. The method of claim 11, wherein receiving the request from the remote website comprises receiving a request from a widget embedded within a web page of the remote website.

15. The method of claim 11, further comprising generating non-personalized web feed recommendations for the user.

16. The method of claim 11, further comprising returning a ranked set of the web feed subscriptions of the user.

17. The method of claim 11, wherein the scores are generated at least partly based on the number of users that selected both of the associated web feeds for subscription or viewing.

18. The method of claim 11, wherein item-to-item mappings between web feeds are stored in an associated items table in computer storage.

19. A computer-implemented method comprising:

by a computer system comprising one or more processors;

maintaining a record of feeds to which a user is subscribed;

maintaining a record of feeds the user has viewed;

detecting behavioral associations between feeds stored in a web feed repository by at least identifying pairs of the stored feeds that were both selected for subscription or viewing by a first plurality of users and creating first item-to-item mappings between the identified stored feeds, the first item-to-item mappings comprising scores reflecting the degree of the associations;

selecting a set of feeds to recommend to the user based, at least in part, on the feeds to which the user has subscribed and the feeds the user has viewed, by at least selecting from the behavioral associations associated stored feeds that are associated with a subset of the feeds to which the user has subscribed and the feeds the user has viewed; and

generating a display page for presentation to the user, the display page including a first area that lists the feeds to which the user is subscribed, a second area that lists the additional feeds selected to recommend, and a third area comprising a text field providing functionality for the user to input a web feed address corresponding to a new web feed, wherein inputting a web feed address by the user makes the new web feed available for other users to select for subscription if the web feed address is not already stored in the web feed repository;

detecting one or more behavioral associations between the new web feed and one or more of the stored feeds based at least partly on behavior of a second plurality of users, wherein said detecting comprises creating second item-to-item mappings between the new web feed and the one or more of the stored web feeds when the new web feed and the one or more of the stored web feeds were both selected for subscription or viewing by the second plurality of users, the second item-to-item mappings comprising second scores reflecting a degree of association between the new web feed and the one or more of the stored web feeds;

recommending the new web feed to one or more other users based at least partly on the one or more behavioral associations.

20. The method of claim 19, wherein generating the display page comprises ranking the feeds to which the user is subscribed based, at least in part, on the following information about each subscribed-to feed:

(a) duration since the feed was last updated, and (b) duration since the user last viewed the feed.

21. The method of claim 19, further comprising selecting a set of feeds to recommend to the user based, at least in part, on data regarding web feed popularity.

22. The method of claim 19, further comprising providing a control on the display page, the control configured to permit the user to contribute a feed to the record of feeds.

23. The method of claim 19, wherein selecting a set of feeds to recommend to the user comprises selecting random web feeds to recommend to the user.

24. The method of claim 19, wherein a web feed address submitted by the user is submitted to the web feed repository only after determining that the web feed address is valid and not already contained in the web feed repository.

25. A computer-implemented method comprising:

by a system comprising computer hardware;

subscribing a first user to a first item in response to the first user contributing the first item;

adding the first item to a data repository in response to the first user contributing the first item;

retrieving subscription data relating to a second user from a user database, the subscription data comprising data regarding subscribed items that the second user has subscribed to;

recommending the first item to the second user, wherein the first item is selected to be recommended to the second user based, at least in part, on a determination that both the first item and at least one of the items the second user has subscribed to were selected for subscription or viewing by a first plurality of users;

making the first item available for the second user to subscribe to by selecting a subscribe control in a user interface;

retrieving the first user'"'"'s subscription data from the user database, the subscription data comprising data regarding subscribed items that the first user has subscribed to;

determining one or more of a last update time of the first user'"'"'s subscribed items and a last viewed time of the first user'"'"'s subscribed items by the first user;

ranking the first user'"'"'s subscribed items based at least in part on the items'"'"' last update time and last viewed time;

selecting a set of second items to recommend to the first user based, at least in part, on the first user'"'"'s subscription data, wherein the second items to recommend are selected by at least identifying pairs of the second items that were both selected for subscription or viewing by a second plurality of other users and creating scored item-to-item mappings between the identified second items and selecting a subset of the associated second items based at least partly on the scores of the associated second items; and

outputting the first user'"'"'s subscribed items and the second item recommendations for display to the first user.

26. The method of claim 25, wherein ranking the subscribed items comprises using at least one exponential decay function to rank the subscribed items.

27. The method of claim 25, further comprising selecting a set of items to recommend to the user.

28. The method of claim 27, wherein the web page is further configured to facilitate selecting one of the recommended items by the user, wherein selecting one of the recommended items subscribes the user to the selected recommended item.

29. The method of claim 27, wherein selecting a set of items to recommend to the user comprises selecting items randomly.