Techniques to facilitate recommendations for non-member connections

1. A method for providing recommended social networking connections, the method comprising:

on a social networking service:

determining a set of connection candidates, using one or more computer processors, based upon information gathered about a member of the social networking service, the connection candidates in the set of connection candidates are not already members of the social networking service;

for each particular connection candidate in the set of connection candidates:

determining a plurality of observed signals that indicate a likelihood that the member knows the particular connection candidate using the one or more computer processors; and

determining a relevance score for the particular connection candidate in the set of connection candidates, using the one or more computer processors, based upon the determined plurality of observed signals; and

presenting to the member the set of connection candidates ordered based upon relevance scores.

2. The method of claim 1 , wherein presenting to the member the set of connection candidates includes presenting only those connection candidates having a relevance score that exceeds a predetermined threshold value.

3. The method of claim 1 , wherein the plurality of observed signals includes at least one of: any communication from a connection candidate in the set of connection candidates to the member; any communication from the member to a connection candidate; the presence of a connection candidate in an address book of the member; an association between an IP address associated with the member and a second IP address associated with a connection candidate; an age similarity between the member and a connection candidate.

4. The method of claim 1 , wherein determining the set of connection candidates based upon information gathered about a member of the social networking service comprises determining a set of connection candidates based upon at least an address book of a connection of the member, and wherein one of the plurality of signals includes the number of connections of the member that have a common connection candidate in their address book.

5. The method of claim 1 , wherein determining the set of connection candidates based upon information gathered about a member of the social networking service comprises determining a set of connection candidates based at least upon contents of emails in an email account of the member.

6. The method of claim 1 , wherein determining the relevance score based upon a plurality of signals comprises determining a relevance score using the plurality of signals as inputs to a Bayesian classifier.

7. The method of claim 1 , wherein determining the relevance score based upon a plurality of signals comprises at least determining a weighted score for each of the plurality of observed signals and determining the relevance score by at least summing the weighted scores for the plurality of observed signals.

8. The method of claim 1 , wherein determining the relevance score based upon a plurality of signals comprises determining a relevance score using at least a neural network.

9. A system for providing recommended social networking connections, the system comprising:

a data collection module configured to:

gather information about a member of the social networking service;

an application module configured to:

determine a set of connection candidates based upon the information gathered about the member of the social networking service, the connection candidates in the set of connection candidates are not already members of the social networking service;

a relevance module configured to:

determine, for each particular connection candidate in the set of connection candidates:

a plurality of observed signals that indicate a likelihood that the member knows the particular connection candidate; and

a relevance score based upon the determined plurality of observed signals; and

the application module is further configured to present to the member the set of connection candidates ordered based upon relevance scores.

10. The system of claim 9 , wherein the application module is configured to present the set of connection candidates by presenting only those connection candidates having a relevance score that exceeds a predetermined threshold value.

11. The system of claim 9 , wherein the plurality of observed signals includes at least one of: any communication from a connection candidate in the set of connection candidates to the member; any communication from the member to a connection candidate; the presence of a connection candidate in an address book of the member; an association between an IP address associated with the member and a second IP address associated with a connection candidate; an age similarity between the member and a connection candidate.

12. The system of claim 9 , wherein the application module is configured to determine the set of connection candidates based upon information gathered about a member of the social networking service by at least determining a set of connection candidates based upon an address book of a connection of the member, and wherein one of the plurality of signals includes the number of connections of the member that have a common connection candidate in their address book.

13. The system of claim 9 , wherein the application module is configured to determine the set of connection candidates based upon information gathered about a member of the social networking service by at least determining a set of connection candidates based upon emails in an email account of the member.

14. The system of claim 9 , wherein the relevance module is configured to determine the relevance score based upon a plurality of observed signals by at least determining a relevance score using the plurality of signals as inputs to a Bayesian classifier.

15. The system of claim 9 , wherein the relevance module is configured to determine the relevance score based upon a plurality of observed signals by at least determining a weighted score for each of the plurality of observed signals and determining the relevance score by summing the weighted scores for the plurality of signals.

16. The system of claim 9 , wherein the relevance module is configured to determine the relevance score based upon a plurality of observed signals by at least determining a relevance score using a neural network.

17. A non-transitory machine readable medium that stores instructions which when performed by a machine, cause the machine to perform operations comprising:

determining a set of connection candidates based upon information gathered about a member of the social networking service, the connection candidates in the set of connection candidates are not already members of the social networking service;

for each particular connection candidate in the set of connection candidates:

determining a plurality of observed signals that indicate a likelihood that the member knows the particular connection candidate using the one or more computer processors; and

determine a relevance score for the particular connection candidate in the set of connection candidates based upon the determined plurality of observed signals; and

presenting to the member the set of connection candidates ordered based upon relevance scores.

18. The machine readable medium of claim 17 , wherein the instructions for presenting to the member the set of connection candidates include instructions, which when performed by the machine, cause the machine to perform the operations of: presenting only those connection candidates having a relevance score that exceeds a predetermined threshold value.

19. The machine readable medium of claim 17 , wherein the plurality of signals includes at least one of: any communication from a connection candidate in the set of connection candidates to the member; any communication from the member to a connection candidate; the presence of a connection candidate in an address book of the member; an association between an IP address associated with the member and a second IP address associated with a connection candidate; an age similarity between the member and a connection candidate.

20. The machine readable medium of claim 17 , wherein the instructions for determining the set of connection candidates based upon information gathered about a member of the social networking service comprise instructions, which when performed by the machine, cause the machine to perform at least the operations of determining a set of connection candidates based upon an address book of a connection of the member, and wherein one of the plurality of signals includes the number of connections of the member that have a common connection candidate in their address book.

21. The machine readable medium of claim 17 , wherein the instructions for determining the set of connection candidates based upon information gathered about a member of the social networking service comprise instructions, which when performed by the machine, cause the machine to at least perform the operations of determining a set of connection candidates based upon contents of emails in an email account of the member.

22. The machine readable medium of claim 17 , wherein the instructions for determining the relevance score based upon a plurality of observed signals comprises instructions, which when performed by the machine, cause the machine to at least perform the operations of determining a relevance score using the plurality of signals as inputs to a Bayesian classifier.

23. The machine readable medium of claim 17 , wherein the instructions for determining the relevance score based upon a plurality of observed signals comprises instructions, which when performed by the machine, cause the machine to at least perform the operations of determining a weighted score for each of the plurality of observed signals and determining the relevance score by summing the weighted scores for the plurality of observed signals.

24. The machine readable medium of claim 17 , wherein the instructions for determining the relevance score based upon a plurality of observed signals comprises instructions, which when performed by the machine, cause the machine to at least perform the operations of determining a relevance score using a neural network.