Query intent clustering for automated sourcing

1. A computer system, comprising:

a processor;

a storage device;

an attribute extractor;

a query generator; and

a memory device holding an instruction set executable on the processor to cause the computer system to perform operations comprising:

retrieving, from a database, one or more search attributes based on a mapping of the one or more search attributes to a given attribute submitted to an online networking service from a client system;

based on one or more search attributes, performing a search to identify one or more recent hire member profiles that contain a search attribute from the one or more search attributes, the one or more recent hire member profiles having been modified with a preset period of time to add the search attribute;

extracting, by the attribute extractor, one or more attributes from the one or more recent hire member profiles;

storing, on the storage device, the extracted one or more attributes;

identifying skills clusters based on the extracted one or more attributes retrieved from the storage device, the skills clusters based at least partially on skills shared by the identified one or more recent hire member profiles as compared to skills shared by member profiles that have not been modified within the preset period of time to add the search attribute; and

generating, by the query generator, a search query based on the identified skills clusters.

2. The computer system of claim 1 , the memory device to perform the operations comprising:

causing presentation of the generated search query on a display device of the client system;

subsequent to the presentation, receiving from the client system, a modified version of the generated search query; and

executing the modified version of the generated search query to retrieve search results.

3. The computer system of claim 1 , wherein identifying the skills clusters comprises:

performing latent Dirichlet allocation (LDA) for the given attribute and one or more search attributes; and

obtaining topic vectors resulting from performing the LDA.

4. The computer system of claim 1 , wherein extracting the one or more attributes comprises:

extracting skills from the one or more recent hire member profiles; and

mapping the extracted skills to skill identifiers.

5. The computer system of claim 4 , wherein identifying the skills clusters comprises performing query intent clustering for the skill identifiers.

6. The computer system of claim 4 , wherein identifying the skills clusters comprises:

creating a skills taxonomy for the skill identifiers; and

representing each skill identifier in the skills taxonomy as a categorical variable that is either absent or present in a member profile.

7. The computer system of claim 4 , wherein identifying the skills clusters comprises:

performing latent class analysis (LCA) to create the skills clusters as clusters of the skill identifiers.

8. A computer-implemented method, comprising:

retrieving, from a database, one or more search attributes based on a mapping of the one or more search attributes to a given attribute submitted to an online networking service from a client system;

based on one or more search attributes, performing a search to identify one or more recent hire member profiles that contain a search attribute from the one or more search attributes, the one or more recent hire member profiles having been modified with a preset period of time to add the search attribute;

extracting, by the attribute extractor, one or more attributes from the one or more recent hire member profiles;

storing, on the storage device, the extracted one or more attributes;

identifying skills clusters based on the extracted one or more attributes retrieved from the storage device, the skills clusters based at least partially on skills shared by the identified one or more recent hire member profiles as compared to skills shared by member profiles that have not been modified within the preset period of time to add the search attribute; and

generating, by the query generator, a search query based on the identified skills clusters.

9. The computer-implemented method of claim 8 , comprising:

causing presentation of the generated search query on a display device of the client system;

subsequent to the presentation, receiving from the client system, a modified version of the generated search query; and

executing the modified version of the generated search query to retrieve search results.

10. The computer-implemented method of claim 8 , wherein identifying the skills clusters comprises:

performing latent Dirichlet allocation (LDA) for the given attribute title and one or more search attributes; and

obtaining topic vectors resulting from performing the LDA.

11. The computer-implemented method of claim 8 , further comprising:

performing the search on the member profiles in the online networking service using the generated search query, returning one or more result member profiles as candidates.

12. The computer-implemented method of claim 8 , wherein extracting the one or more attributes comprises:

extracting skills from the one or more recent hire member profiles; and

mapping the extracted skills to skill identifiers.

13. The computer-implemented method of claim 12 , wherein identifying the skills clusters comprises performing query intent clustering for the skill identifiers.

14. The computer-implemented method of claim 12 , wherein identifying the skills clusters comprises:

creating a skills taxonomy for the skill identifiers; and

representing each skill identifier in the skills taxonomy as a categorical variable that is either absent or present in a member profile.

15. The computer-implemented method of claim 12 , wherein identifying the skills clusters comprises:

performing latent class analysis (LCA) to create the skills clusters as clusters of the skill identifiers.

16. A non-transitory machine-readable storage medium comprising instructions that, when executed by one or more processors of a machine, cause the machine to:

retrieving, from a database, one or more search attributes based on a mapping of the one or more search attributes to a given attribute submitted to an online networking service from a client system;

based on one or more search attributes, performing a search to identify one or more recent hire member profiles that contain a search attribute from the one or more search attributes, the one or more recent hire member profiles having been modified with a preset period of time to add the search attribute;

extracting, by the attribute extractor, one or more attributes from the one or more recent hire member profiles;

storing, on the storage device, the extracted one or more attributes;

identifying skills clusters based on the extracted one or more attributes retrieved from the storage device, the skills clusters based at least partially on skills shared by the identified one or more recent hire member profiles as compared to skills shared by member profiles that have not been modified within the preset period of time to add the search attribute; and

generating, by the query generator, a search query based on the identified skills clusters.