MEMBER FEATURE SETS, GROUP FEATURE SETS AND TRAINED COEFFICIENTS FOR RECOMMENDING RELEVANT GROUPS
A system, a machine-readable storage medium storing instructions, and a computer-implemented method are described herein to a Group Relevance Engine that generates, for a group in a social network, an aggregate group feature based on a common attribute shared amongst member accounts currently subscribed to the group. The Group Relevance Engine identifies an account feature corresponding to the common attribute in a profile of a target member account. The Group Relevance calculates a relevance score based at least on a match between the aggregate group feature and the account feature. The Group Relevance determines whether to recommend the group to the target member account based at least on the relevance score.
1 . A computer system comprising:
a processor;
a memory device holding an instruction set executable on the processor to cause the computer system to perform operations comprising:
generating, for a group in a social network, an aggregate group feature based on a common attribute shared amongst member accounts currently subscribed to the group;
identifying an account feature corresponding to the common attribute in a profile of a target member account;
calculating a relevance score based at least on a match between the aggregate group feature and the account feature; and
determining whether to recommend the group to the target member account based at least on the relevance score.
2 . The computer system of claim 1 , wherein generating, for a group in a social network, an aggregate group feature based on a common attribute shared amongst member accounts currently subscribed to the group comprises:
identifying a percentage of member accounts currently subscribed to the group that each have a common type of profile attribute;
determining the percentage of member accounts meets a percentage threshold; and
setting the aggregate group feature to the common type of profile attribute.
3 . The computer system of claim 2 , wherein the common type of profile attribute comprises at least one of: current industry, gender, professional experience, country, geographical region, educational degree, occupational function, employer and skill.
4 . The computer system of claim 1 , wherein calculating a relevance score based at least on a match between the aggregate group feature and the account feature comprises:
determining the match between the aggregate group feature and the account feature;
identifying an updateable learned coefficient corresponding to the match of the aggregate group feature and the account feature; and
identifying a value of the account feature;
identifying a value of the aggregate group feature; and
calculating the relevance score based at least on the value of the account feature, the value of the aggregate group feature and the updateable learned coefficient.
5 . The computer system of claim 4 , wherein the updateable learned coefficient represents a learned weighting of importance of the match in calculating the relevance score.
6 . The computer system of claim 4 , wherein the value of the aggregate group feature is based on a percentage of member accounts currently subscribed to the group that each have a type of profile attribute used as the aggregate group feature.
7 . The computer system of claim 4 , wherein determining the match between the aggregate group feature and the account feature comprises:
determining the match based on a cosine similarity between the aggregate group feature and the account feature.
8 . A computer-implemented method comprising:
generating, for a group in a social network, an aggregate group feature based on a common attribute shared amongst member accounts currently subscribed to the group;
identifying an account feature corresponding to the common attribute in a profile of a target member account;
calculating, using one or more processors, a relevance score based at least on a match between the aggregate group feature and the account feature; and
determining whether to recommend the group to the target member account based at least on the relevance score.
9 . The computer-implemented method of claim 8 , wherein generating, for a group in a social network, an aggregate group feature based on a common attribute shared amongst member accounts currently subscribed to the group comprises:
identifying a percentage of member accounts currently subscribed to the group that each have a common type of profile attribute;
determining the percentage of member accounts meets a percentage threshold; and
setting the aggregate group feature to the common type of profile attribute.
10 . The computer-implemented method of claim 9 , wherein the common type of profile attribute comprises at least one of: current industry, gender, professional experience, country, geographical region, educational degree, occupational function, employer and skill.
11 . The computer-implemented method of claim 8 , wherein calculating a relevance score based at least on a match between the aggregate group feature and the account feature comprises:
determining the match between the aggregate group feature and the account feature;
identifying an updateable learned coefficient corresponding to the match of the aggregate group feature and the account feature; and
identifying a value of the account feature;
identifying a value of the aggregate group feature; and
calculating the relevance score based at least on the value of the account feature, the value of the aggregate group feature and the updateable learned coefficient.
12 . The computer-implemented method of claim 11 , wherein the updateable learned coefficient represents a learned weighting of importance of the match in calculating the relevance score.
13 . The computer-implemented method of claim 11 , wherein the value of the aggregate group feature is based on a percentage of member accounts currently subscribed to the group that each have a type of profile attribute used as the aggregate group feature.
14 . A non-transitory computer-readable medium storing executable instructions thereon, which, when executed by a processor, cause the processor to perform operations including:
generating, for a group in a social network, an aggregate group feature based on a common attribute shared amongst member accounts currently subscribed to the group;
identifying an account feature corresponding to the common attribute in a profile of a target member account;
calculating a relevance score based at least on a match between the aggregate group feature and the account feature; and
determining whether to recommend the group to the target member account based at least on the relevance score.
15 . The non-transitory computer-readable medium of claim 14 , wherein generating, for a group in a social network, an aggregate group feature based on a common attribute shared amongst member accounts currently subscribed to the group comprises:
identifying a percentage of member accounts currently subscribed to the group that each have a common type of profile attribute;
determining the percentage of member accounts meets a percentage threshold; and
setting the aggregate group feature to the common type of profile attribute.
16 . The non-transitory computer-readable medium of claim 15 , wherein the common type of profile attribute comprises at least one of: current industry, gender, professional experience, country, geographical region, educational degree, occupational function, employer and skill.
17 . The non-transitory computer-readable medium of claim 14 , wherein calculating a relevance score based at least on a match between the aggregate group feature and the account feature comprises:
determining the match between the aggregate group feature and the account feature;
identifying an updateable learned coefficient corresponding to the match of the aggregate group feature and the account feature; and
identifying a value of the account feature;
identifying a value of the aggregate group feature; and
calculating the relevance score based at least on the value of the account feature, the value of the aggregate group feature and the updateable learned coefficient.
18 . The non-transitory computer-readable medium of claim 17 , wherein the updateable learned coefficient represents a learned weighting of importance of the match in calculating the relevance score.
19 . The non-transitory computer-readable medium of claim 17 , wherein the value of the aggregate group feature is based on a percentage of member accounts currently subscribed to the group that each have a type of profile attribute used as the aggregate group feature.
20 . The non-transitory computer-readable medium of claim 17 , wherein determining the match between the aggregate group feature and the account feature comprises:
determining the match based on a cosine similarity between the aggregate group feature and the account feature.