Recommendations for online system groups

Patent metadata

Field	Value
Publication number	US-10877976-B2
Application number	US-201715585628-A
Country	US
Kind code	B2
Filing date	May 3, 2017
Priority date	May 3, 2017
Publication date	Dec 29, 2020
Grant date	Dec 29, 2020

Abstract

An online system provides group recommendations by applying a set of sourcing rules to identify a plurality of candidate groups and then generating scores for the candidate groups. The sourcing rules can be configured to identify a relatively small subset of the groups maintained by the online system. After the candidate groups are identified, the online system generates a score for each candidate group, ranks the candidate groups based on the scores, and sends high-ranking candidate groups to the target user to be displayed as recommended groups. As a result, the online system generates a smaller number of scores, which advantageously allows for the online system to provide group recommendations to users in a more computationally efficient manner.

First claim

What is claimed is:

1. A method comprising:
maintaining a user account for each of a plurality of users of an online system;
maintaining a plurality of groups in the online system, each group enabling the users of the online system to interact with other users who are members of the group;
identifying a target user of the plurality of users to recommend to join one or more selected groups of the plurality of groups;
selecting a plurality of groups as candidate groups for recommending to the target user, the selection of the plurality of groups based on a plurality of sourcing rules;
for each of the candidate groups, computing a score for the candidate group representing a likelihood of the target user joining the group if a recommendation to join the group is displayed to the target user by:
providing a set of features extracted from the candidate group as input to a machine learning model trained based on a training set comprising groups that the target user chose to join after being presented with a recommendation to join the group;

selecting one or more of the candidate groups for display to the viewing user, the selection of a candidate group based on the score for the candidate group; and
sending a recommendation to join the selected groups for display to the target user.

2. The method of claim 1 , wherein one of the sourcing rules identifies a group as a candidate group if a user connected to the target user is a member of the group.

3. The method of claim 1 , wherein one of the sourcing rules identifies a group as a candidate group if the group is associated with a location less than a threshold distance from a location of the target user.

4. The method of claim 1 , wherein one of the sourcing rules identifies a group as a candidate group if the group has an activity level exceeding a threshold activity level.

5. The method of claim 1 , wherein one of the sourcing rules identifies a group as a candidate group if the group is associated with one or more attributes that match attributes of the target user.

6. The method of claim 1 , wherein the selected groups are displayed in a user interface in a plurality of categories, wherein the selected groups displayed in each category were identified as candidate groups by a sourcing rule corresponding to the category.

7. The method of claim 1 , wherein each of the plurality of groups in the online system is associated with one or more topics representing the subject matter of the group, and wherein each selected group is displayed in a category matching a topic associated with the selected group.

8. The method of claim 6 , further comprising:
for each of the plurality of categories, computing a score for the category, the score representing a prediction of the target user's interest in the category;
determining an ordering for the plurality of categories based on the score for each of the plurality of categories,
wherein the plurality of categories are displayed in the user interface in the determined ordering.

9. The method of claim 8 , wherein computing the score for a category comprises determining the target user's interest in the category based on connections between the target user and a plurality of brand pages and groups related to the category.

10. The method of claim 8 , wherein the score for a category is computed based on an average score for groups in the category.

11. The method of claim 8 , wherein the score for a category is computed based on the score for the group having the highest score of the groups in the category.

12. The method of claim 1 , wherein the selection of a candidate group is further based on one or more diversity rules.

13. A non-transitory computer readable storage medium comprising instructions which, when executed by a processor, cause the processor to perform the steps of:
maintaining a user account for each of a plurality of users of an online system;
maintaining a plurality of groups in the online system, each group enabling the users of the online system to interact with other users who are members of the group;
identifying a target user of the plurality of users to recommend to join one or more selected groups of the plurality of groups;
selecting a plurality of groups as candidate groups for recommending to the target user, the selection of the plurality of groups based on a plurality of sourcing rules;
for each of the candidate groups, computing a score for the candidate group representing a likelihood of the target user joining the group if a recommendation to join the group is displayed to the target user by:
providing a set of features extracted from the candidate group as input to a machine learning model trained based on a training set comprising groups that the target user chose to join after being presented with a recommendation to join the group;

selecting one or more of the candidate groups for display to the viewing user, the selection of a candidate group based on the score for the candidate group; and
sending a recommendation to join the selected groups for display to the target user.

14. The non-transitory computer readable storage medium of claim 13 , wherein one of the sourcing rules identifies a group as a candidate group if a user connected to the target user is a member of the group.

15. The non-transitory computer readable storage medium of claim 13 , wherein one of the sourcing rules identifies a group as a candidate group if the group is associated with a location less than a threshold distance from a location of the target user.

16. The non-transitory computer readable storage medium of claim 13 , wherein one of the sourcing rules identifies a group as a candidate group if the group has an activity level exceeding a threshold activity level.

17. The non-transitory computer readable storage medium of claim 13 , wherein one of the sourcing rules identifies a group as a candidate group if the group is associated with one or more attributes that match attributes of the target user.

18. The non-transitory computer readable storage medium of claim 13 , wherein the selected groups are displayed in a user interface in a plurality of categories, wherein the selected groups displayed in each category were identified as candidate groups by a sourcing rule corresponding to the category.

19. The non-transitory computer readable storage medium of claim 13 , wherein each of the plurality of groups in the online system is associated with one or more topics representing the subject matter of the group, and wherein each selected group is displayed in a category matching a topic associated with the selected group.

20. The non-transitory computer readable storage medium of claim 13 , wherein the selection of a candidate group is further based on one or more diversity rules.

Assignees

• Facebook Inc

Inventors

• Miao Yi
• Mayes Kelly Jayne
• Ertmann Eric Carl
• Ju Li
• Zhang Congle

Classifications

• G06Q10/40

  Business processes related to social networking or social networking services · CPC title

• H04L67/52

  specially adapted for the location of the user terminal · CPC title

• G06N5/046 Primary

  Forward inferencing; Production systems · CPC title

• G06Q30/0631

  Recommending goods or services · CPC title

• G06N20/00

  Machine learning · CPC title