JOINT OPTIMIZATION AND ASSIGNMENT OF JOB RECOMMENDATIONS
An on-line social network system includes or is in communication with a recommendation system that is configured to assign jobs to members while taking into account the relevance of a job for a given member as well as the relevance of the same job for other members. The objective of said optimization is to maximize the total sum of respective relevance scores generated for member/job pairs for jobs that get selected for presentation to members. The optimization objective is constrained by the maximum number of job recommendations desirable for each member profile and may also be constrained by the maximum number of member recommendations desirable for each job posting.
1 . A computer-implemented method comprising:
accessing a subject member profile representing a member in an on-line social network system, the subject member profile is from a set of member profiles;
using at least one processor, selecting a set of candidate job postings for the subject member profile, the set of candidate job postings is from a larger set of job postings;
for each job posting in the set of candidate job postings, generating an associated relevance value indicating a likelihood that a member represented by the subject member profile applies for a job represented by the job posting;
from the set of candidate job postings, using at least one processor, selecting a presentation set based on:
respective relevance values generated for each job posting from the larger set of job postings with respect to each member profile from the set of member profiles, and
a predetermined jobs cap value indicating the maximum number of job postings that is to be selected for presentation to a member represented by a profile from the set of member profiles; and
causing presentation, on a display device, of a reference to a job posting from the presentation set.
2 . The method of claim 1 , wherein the selecting of the presentation set comprises:
generating, for each pair comprising a member profile from the set of member profiles and a job posting from the larger set of job postings, a binary value indicating whether the associated job posting is assigned to the associated member profile; and
including, in the presentation set, those job postings, for which the binary value with respect to the subject member profile is indicative of positive assignment.
3 . The method of claim 2 , wherein the generating of respective binary values for each pair comprising a member profile from the set of member profiles and a job posting from the larger set of job postings indicating whether the associated job posting is assigned to the associated member comprises:
constructing a weighted bipartite graph with nodes representing member profiles from the set of member profiles and job postings from the larger set of job postings, an edge associated with a node representing a member profile and a node representing a job posting having a weight reflecting a relevance value generated for a par comprising that member profile and that job posting; and
calculating a maximum weighted bipartite matching with respect to the constructed weighted bipartite graph.
4 . The method of claim 3 , wherein a number of nodes representing the subject member profile in the constructed weighted bipartite graph equals the predetermined jobs cap value.
5 . The method of claim 3 , wherein a number of nodes representing a particular job posting in the constructed weighted bipartite graph equals a predetermined candidates cap value.
6 . The method of claim 1 , wherein the selecting of the presentation set comprises utilizing greedy algorithm with respect to a set of relevance values generated for each pair comprising a member profile from the set of member profiles and a job posting from the larger set of job postings.
7 . The method of claim 1 , wherein each member profile in the set of member profiles is associated with a particular industry.
8 . The method of claim 1 , wherein each job posting in the larger set of job postings is associated with a particular geographic location.
9 . The method of claim 1 , wherein each member profiles from the set of member wherein profiles is associated with a profile UI configured to display and to collect data about the associated member.
10 . The method of claim 1 , wherein each job posting in the larger set of job postings is an electronic publication in a structured form.
11 . A computer-implemented system comprising:
an access module, implemented using at least one processor, to access a subject member profile representing a member in an on-line social network system, the subject member profile is from a set of member profiles;
a preliminary set selector, implemented using at least one processor, to select a set of candidate job postings for the subject member profile, the set of candidate job postings is from a larger set of job postings;
a relevance value generator, implemented using at least one processor, to generate, for each job posting in the set of candidate job postings, an associated relevance value indicating a likelihood that a member represented by the subject member profile applies for a job represented by the job posting;
a presentation set selector, implemented using at least one processor, to select from the set of candidate job postings, a presentation set based on:
respective relevance values generated for each job posting from the larger set of job postings with respect to each member profile from the set of member profiles, and
a predetermined jobs cap value indicating the maximum number of job postings that is to be selected for presentation to a member represented by a profile from the set of member profiles; and
a presentation module, implemented using at least one processor, to cause presentation, on a display device, of a reference to a job posting from the presentation set.
12 . The system of claim 11 , wherein the presentation set selector is to:
generate, for each pair comprising a member profile from the set of member profiles and a job posting from the larger set of job postings, a binary value indicating whether the associated job posting is assigned to the associated member profile; and
include, in the presentation set, those job postings, for which the binary value with respect to the subject member profile is indicative of positive assignment.
13 . The system of claim 12 , wherein the generating of respective binary values for each pair comprising a member profile from the set of member profiles and a job posting from the larger set of job postings indicating whether the associated job posting is assigned to the associated member comprises:
constructing a weighted bipartite graph with nodes representing member profiles from the set of member profiles and job postings from the larger set of job postings, an edge associated with a node representing a member profile and a node representing a job posting having a weight reflecting a relevance value generated for a par comprising that member profile and that job posting; and
calculating a maximum weighted bipartite matching with respect to the constructed weighted bipartite graph.
14 . The system of claim 13 , wherein a number of nodes representing the subject member profile in the constructed weighted bipartite graph equals the predetermined jobs cap value.
15 . The system of claim 13 , wherein a number of nodes representing a particular job posting in the constructed weighted bipartite graph equals a predetermined candidates cap value.
16 . The system of claim 11 , wherein the selecting of the presentation set comprises utilizing greedy algorithm with respect to a set of relevance values generated for each pair comprising a member profile from the set of member profiles and a job posting from the larger set of job postings.
17 . The system of claim 11 , wherein each member profile in the set of member profiles is associated with a particular industry.
18 . The system of claim 11 , wherein each job posting in the larger set of job postings is associated with a particular geographic location.
19 . The system of claim 11 , wherein each member profiles from the set of member wherein profiles is associated with a profile UI configured to display and to collect data about the associated member.
20 . A machine-readable non-transitory storage medium having instruction data executable by a machine to cause the machine to perform operations comprising:
accessing a subject member profile representing a member in an on-line social network system, the subject member profile is from a set of member profiles;
using at least one processor, selecting a set of candidate job postings for the subject member profile, the set of candidate job postings is from a larger set of job postings;
for each job posting in the set of candidate job postings, generating an associated relevance value indicating a likelihood that a member represented by the subject member profile applies for a job represented by the job posting;
from the set of candidate job postings selecting a presentation set based on:
respective relevance values generated for each job posting from the larger set of job postings with respect to each member profile from the set of member profiles, and
a predetermined jobs cap value indicating the maximum number of job postings that is to be selected for presentation to a member represented by a profile from the set of member profiles; and
causing presentation, on a display device, of a reference to a job posting from the presentation set.