Entity-aware features for personalized job search ranking

1. A computer-implemented method for automatically augmenting digitally stored member profiles with skills the members have not explicitly added to their member profiles, the method comprising:

obtaining a plurality of member profiles in a social networking service, each member profile identifying a member and listing one or more skills the corresponding member has explicitly added to the member profile, the one or more skills indicating a proficiency by the member in the corresponding skill;

forming a members-skills matrix, wherein each cell in the matrix is assigned a value based on whether the corresponding member has the corresponding skill;

computing the dot product of the members matrix and the skills matrix;

using the dot product to identify one or more latent skills of a first member of the social networking service;

augmenting a first digitally stored member profile for the first member with the one or more latent skills by combining the one or more latent skills with explicit skills listed in the first digitally stored member profile for purposes of one or more searches that utilize member skills as an input variable; and

returning results from the one or more searches, the results based on the explicit skills listed in the first digitally stored member profile and the one or more latent skills of the first member.

2. The method of claim 1 , further comprising:

receiving a request for one or more job postings relevant to a first member of the social networking service;

using the dot product of the members matrix and the skills matrix to identify one or more latent skills the first member is likely to have despite the first member not explicitly listing the one or more latent skills in a corresponding member profile; and

using the identified one or more latent skills to rank job postings relevant to the first member of the social networking service.

3. The method of claim 2 , wherein the request includes a search term query from the member; and

the one or more job postings relevant to the first member of the social networking service include job postings containing the one or more search terms.

4. The method of claim 3 , wherein job postings are indexed in a job posting database using terms extracted from the job postings and assigned a plurality of different entity types.

5. The method of claim 4 , further comprising:

segmenting the search term query into a plurality of segments, wherein at least one of the plurality of segments is tagged as a first entity type and at least one of the plurality of segments is tagged as a second entity type; and

wherein the tagged segments are used to retrieve relevant job postings from the job posting database by comparing the tagged segments to index terms of the same entity types.

6. The method of claim 1 , wherein the members matrix includes, for each member in the members-skill matrix, a row containing values indicating a probability that the corresponding member has skills in a corresponding one of k latent topic groupings of skills.

7. The method of claim 1 , wherein the skills matrix includes, for each skill in the members-skill matrix, a column containing values indicating a probability that the corresponding skill applies to a corresponding one of k latent topic groupings of skills.

8. A system comprising:

a computer readable medium having instructions stored there on, which, when executed by a processor, cause the system to:

obtain a plurality of member profiles in a social networking service, each member profile identifying a member and listing one or more skills the corresponding member has explicitly added to the member profile, the one or more skills indicating a proficiency by the member in the corresponding skill;

form a members-skills matrix, wherein each cell in the matrix is assigned a value based on whether the corresponding member has the corresponding skill;

compute the dot product of the members matrix and the skills matrix;

compute the dot product of the members matrix and the skills matrix;

use the dot product to identify one or more latent skills of a first member of the social networking service;

augment a first digitally stored member profile for the first member with the one or more latent skills by combining the one or more latent skills with explicit skills listed in the first digitally, stored member profile for purposes of one or more searches that utilize member skills as an input variable; and

return results from the one or more searches, the results based on the explicit skills listed in the first digitally stored member profile and the one or more latent skills of the first member.

9. The system of claim 8 , wherein the computer readable medium further has instructions that cause the system to:

receive a request for one or more job postings relevant to a first member of the social networking service;

use the dot product of the members matrix and the skills matrix to identify one or more latent skills the first member is likely to have despite the first member not explicitly listing the one or more latent skills in a corresponding member profile; and

use the identified one or more latent skills to rank job postings relevant to the first member of the social networking service.

10. The system of claim 9 , wherein the request includes a search term query from the member; and

the one or more job postings relevant to the first member of the social networking service include job postings containing the one or more search terms.

11. The system of claim 10 , wherein job postings are indexed in a job posting database using terms extracted from the job postings and assigned a plurality of different entity types.

12. The system of claim 11 , wherein the computer readable medium further has instructions that cause the system to:

segment the search term query into a plurality of segments, wherein at least one of the plurality of segments is tagged as a first entity type and at least one of the plurality of segments is tagged as a second entity type; and

wherein the tagged segments are used to retrieve relevant job postings from the job posting database by comparing the tagged segments to index terms of the same entity types.

13. The system of claim 8 , wherein the members matrix includes, for each member in the members-skill matrix, a row containing values indicating a probability that the corresponding member has skills in a corresponding one of k latent topic groupings of skills.

14. The system of claim 8 , wherein the skills matrix includes, for each skill in the members-skill matrix, a column containing values indicating a probability that the corresponding skill applies to a corresponding one of k latent topic groupings of skills.

15. A non-transitory machine-readable storage medium comprising instructions, which when implemented by one or more machines, cause the one or more machines to perform operations comprising:

obtaining a plurality of member profiles in a social networking service, each member profile identifying a member and listing one or more skills the corresponding member has explicitly added to the member profile, the one or more skills indicating a proficiency by the member in the corresponding skill;

forming a members-skills matrix, wherein each cell in the matrix is assigned a value based on whether the corresponding member has the corresponding skill;

computing the dot product of the members matrix and the skills matrix;

computing the dot product of the members matrix and the skills matrix;

using the dot product to identify one or more latent skills of a first member of the social networking service;

augmenting a first digitally stored member profile for the first member with the one or more latent skills by combining the one or more latent skills with explicit skills listed in the first digitally stored member profile for purposes of one or more searches that utilize member skills as an input variable; and

returning results from the one or more searches, the results based on the explicit skills listed in the first digitally stored member profile and the one or more latent skills of the first member.

16. The non-transitory machine-readable storage medium of claim 15 , further comprising:

receiving a request for one or more job postings relevant to a first member of the social networking service;

using the dot product of the members matrix and the skills matrix to identify one or more latent skills the first member is likely to have despite the first member not explicitly listing the one or more latent skills in a corresponding member profile; and

using the identified one or more latent skills to rank job postings relevant to the first member of the social networking service.

17. The non-transitory machine-readable storage medium of claim 16 , wherein the request includes a search term query from the member; and

the one or more job postings relevant to the first member of the social networking service include job postings containing the one or more search terms.

18. The non-transitory machine-readable storage medium of claim 17 , wherein job postings are indexed in a job posting database using terms extracted from the job postings and assigned a plurality of different entity types.

19. The non-transitory machine-readable storage medium of claim 18 , further comprising:

segmenting the search term query into a plurality of segments, wherein at least one of the plurality of segments is tagged as a first entity type and at least one of the plurality of segments is tagged as a second entity type; and

wherein the tagged segments are used to retrieve relevant job postings from the job posting database by comparing the tagged segments to index terms of the same entity types.

20. The non-transitory machine-readable storage medium of claim 15 , wherein the members matrix includes, for each member in the members-skill matrix, a row containing values indicating a probability that the corresponding member has skills in a corresponding one of k latent topic groupings of skills.