Recommending objects to a user of a social networking system based on implicit interactions between the user and the recommended objects

1. A method comprising:

obtaining interaction information identifying an implicit interaction between a user and a plurality of objects of a social networking system with which the user has not established a connection;

identifying, based on the obtained interaction information, a set of soft connections between the user and the plurality of objects with which the user has interacted implicitly;

generating a candidate list of soft connections from the set of soft connections;

identifying additional objects to include as soft connections in the candidate list of soft connections based on a similarity between the additional objects and objects of the soft connections in the candidate list;

extracting ranking signals from the soft connections in the candidate list of soft connections;

ranking the soft connections in the candidate list of soft connections based on the extracted ranking signals; and

selecting one or more objects identified by the soft connections from the candidate list of soft connections to recommend to the user based on the rank associated with the soft connection.

2. The method of claim 1 , wherein identifying a set of soft connections comprises identifying a soft connection between an object and the user based on identifying more than a threshold number of implicit interactions between the user and the object.

3. The method of claim 1 , wherein identifying a set of soft connections comprises identifying a soft connection between an object and the user based on an implicit interaction between the user and content associated with the object.

4. The method of claim 1 , wherein identifying a set of soft connections comprises:

identifying an implicit interaction between the user and a topic;

identifying an object associated with the topic from a cluster of pages associated with the topic; and

identifying a soft connection between the user and the object associated with the topic.

5. The method of claim 1 , wherein generating a candidate list of soft connections from the set of soft connections comprises:

identifying a graph distance between the object identified by a soft connection in the set of soft connections and an object with which the user has established a connection;

placing the soft connection in the candidate list of soft connections on determining that the graph distance is less than a threshold value.

6. The method of claim 5 , wherein the graph distance is weighted based on the type of object with which the user has established a connection or the type of object identified as a soft connection to the user.

7. The method of claim 1 , wherein generating a candidate list of soft connections from the set of soft connections comprises:

identifying soft connections from the set of soft connections associated with objects with which the user has interacted with within a time interval;

determining a weight associated with the soft connections from the set of soft connections based on the type of interactions between the user and the objects identified by the soft connections; and

placing soft connections in the candidate list of soft connections based on determining that the weight associated with a soft connection is greater than a threshold value.

8. The method of claim 1 , wherein generating a candidate list of soft connections from the set of soft connections comprises:

identifying soft connections associated with objects with which the user has interacted with for greater than a threshold time interval;

placing the identified soft connections in the candidate list of soft connections.

9. The method of claim 1 , wherein identifying additional objects to include as soft connections in the candidate list of soft connections comprises:

identifying, as additional objects, objects connected to a topic associated with an object identified by a soft connection in the candidate list of soft connections; and

placing the identified objects, as soft connections, in the candidate list of soft connections to recommend to the user.

10. The method of claim 1 , wherein identifying additional objects to include as soft connections in the candidate list of soft connections comprises:

identifying, as additional objects, objects connected to an object identified by a soft connection in the candidate list of soft connections with greater than a threshold number of user of the social networking system; and

placing the identified objects, as soft connections, in the candidate list of soft connections to recommend to the user.

11. The method of claim 1 , wherein the extracted signals are selected from a group comprising a time interval since the last interaction between the user and the object identified by a soft connection in the candidate list of soft connections, a frequency at which the user interacts with the same object identified by a soft connection in the candidate list of soft connections over a threshold period of time, or the total number of types of interactions between the user and the object identified by the soft connection in the candidate list of soft connections.

12. A computer program product comprising a non-transitory computer-readable storage medium containing computer program code for:

obtaining interaction information identifying an implicit interaction between a user and a plurality of objects of a social networking system with which the user has not established a connection;

identifying, based on the obtained interaction information, a set of soft connections between the user and the plurality of objects with which the user has interacted implicitly;

generating a candidate list of soft connections from the set of soft connections;

identifying additional objects to include as soft connections in the candidate list of soft connections based on a similarity between the additional objects and objects of the soft connections in the candidate list;

extracting ranking signals from the soft connections in the candidate list of soft connections;

ranking the soft connections in the candidate list of soft connections based on the extracted ranking signals; and

selecting one or more objects of the soft connections from the candidate list of soft connections to recommend to the user based on the rank associated with the soft connection.

13. The computer program product of claim 12 , wherein identifying a set of soft connections comprises identifying a soft connection between an object and the user on identifying more than a threshold number of implicit interactions between the user and the object.

14. The computer program product of claim 12 , wherein identifying a set of soft connections comprises identifying a soft connection between an object and the user based on an implicit interaction between the user and content associated with the object.

15. The computer program product of claim 12 , wherein identifying a set of soft connections comprises:

identifying an implicit interaction between the user and a topic;

identifying an object associated with the topic from a cluster of pages associated with the topic; and

identifying a soft connection between the user and the object associated with the topic.

16. The computer program product of claim 12 , wherein generating a candidate list of soft connections from the set of soft connections comprises:

identifying a graph distance between the object identified by a soft connection in the set of soft connections and an object with which the user has established a connection;

placing the soft connection in the candidate list of soft connections on determining that the graph distance is less than a threshold value.

17. The computer program product of claim 16 , wherein the graph distance is weighted based on the type of object with which the user has established a connection or the type of object identified as a soft connection to the user.

18. The computer program product of claim 12 , wherein generating a candidate list of soft connections from the set of soft connections comprises:

identifying soft connections from the set of soft connections associated with objects with which the user has interacted with within a time interval;

determining a weight associated with the soft connections from the set of soft connections based on the type of interactions between the user and the objects identified by the soft connections; and

placing soft connections in the candidate list of soft connections on determining that the weight associated with a soft connection is greater than a threshold value.

19. The computer program product of claim 12 , wherein generating a candidate list of soft connections from the set of soft connections comprises:

identifying soft connections associated with objects with which the user has interacted with for greater than a threshold time interval;

placing the identified soft connections in the candidate list of soft connections.

20. The computer program product of claim 12 , wherein identifying additional objects to include as soft connections in the candidate list of soft connections based on one or more soft connections in the candidate list of soft connections comprises:

identifying, as additional objects, objects connected to a topic associated with an object identified by a soft connection in the candidate list of soft connections; and

placing the identified objects, as soft connections, in the candidate list of soft connections to recommend to the user.

21. The computer program product of claim 12 , wherein identifying additional objects to include as soft connections in the candidate list of soft connections based on one or more soft connections in the candidate list of soft connections comprises:

identifying, as additional objects, objects connected to an object identified by a soft connection in the candidate list of soft connections with greater than a threshold number of user of the social networking system; and

placing the identified objects in the candidate list of soft connections to recommend to the user.

22. The computer program product of claim 12 , wherein the extracted signals are selected from a group comprising a time interval since the last interaction between the user and the object identified by a soft connection in the candidate list of soft connections, a frequency at which the user interacts with the same object identified by a soft connection in the candidate list of soft connections over a threshold period of time, or the total number of types of interactions between the user and the object identified by the soft connection in the candidate list of soft connections.