Job skill taxonomy

1. A method for classification of job skills, comprising:

identifying a set of job skill terms from a corpus of job description documents;

generating a synonym file by associating each of the job skill terms with a canonical job skill from a predetermined controlled vocabulary of job skills, wherein at least one canonical job skill is associated with a plurality of job skill terms;

extracting a set of job skills from the corpus of job description documents using the synonym file by identifying skill variants based on the synonym file, wrapping text tokens corresponding to the skill variants in a skill tag, removing non-skill terms, and standardizing the skill variants;

generating job skill relationship information including a parent-child relationship between a parent skill and a child skill based on the extracted set of job skills by performing latent semantic analysis (LSA) on the corpus of job description documents, wherein the job skill relationship information includes skill pairs sorted by a ranking score;

creating a job skill taxonomy based on the job skill relationship information, wherein the job skill taxonomy comprises a hierarchy of job skills including the parent skill and the child skill of the parent skill;

identifying profile information for a job candidate;

identifying the child skill as a missing skill for the job candidate based on the profile information and the parent child relationship; and

generating a career trajectory recommendation for the job candidate including a recommendation to obtain the missing skill.

2. The method of claim 1 , further comprising:

identifying profile information for a job candidate; and

generating a salary prediction based on the profile information and the job skill taxonomy.

3. The method of claim 1 , further comprising: identifying one or more intermediate positions, wherein the career trajectory recommendation includes a recommendation to seek a job related to the one or more intermediate positions.

4. The method of claim 1 , further comprising:

identifying profile information for a job candidate;

identifying a plurality of job postings, wherein each of the plurality of job postings comprises a job description;

scoring each of the plurality of job postings based on the profile information, the corresponding job description, and the job skill taxonomy; and

presenting one or more of the plurality of job postings to the job candidate based on the scoring.

5. The method of claim 1 , further comprising:

identifying a job description for a job posting;

identifying a plurality of candidate resumes;

scoring each of the plurality of candidate resumes based on the job description and the job skill taxonomy; and

presenting one or more of the candidate resumes based on the scoring.

6. The method of claim 1 , further comprising:

receiving a plurality of resumes from a job matching service, wherein the corpus of job description documents comprises the plurality of resumes.

7. The method of claim 1 , further comprising:

receiving a plurality of job descriptions from a job matching service, wherein the corpus of job description documents comprises the plurality of job descriptions.

8. The method of claim 1 , wherein:

the job skill taxonomy comprises a plurality of job skills, wherein each of the job skills is associated with a skill name, a skill description, and one or more related skills.

9. A method for generating a controlled vocabulary of job skills, comprising:

identify a database of resumes, a set of job postings, or both;

extract raw skills from the set of resumes, the set of job postings, or both;

validate the extracted raw skills;

generate a controlled job skill vocabulary based on the validated raw skills;

identify relationship information including a parent-child relationship between a parent skill and a child skill among jobs based on the controlled job skill vocabulary by performing latent semantic analysis (LSA) on the corpus of job description documents, wherein the job skill relationship information includes skill pairs sorted by a ranking score;

creating a job skill taxonomy based on the relationship information, wherein the job skill taxonomy comprises a hierarchy of job skills including the parent skill and the child skill of the parent skill;

identifying profile information for a job candidate;

identifying the child skill as a missing skill for the job candidate based on the profile information and the parent child relationship; and

generating a career trajectory recommendation for the job candidate including a recommendation to obtain the missing skill.

10. The method of claim 9 , further comprising:

extract a set of standardized job skills from the set of resumes, the set of job postings, or both based on the controlled job skill vocabulary.

11. The method of claim 9 , wherein:

the extracted raw skills are validated using an automated algorithm, human validation, or both.

12. The method of claim 9 , wherein:

the extracted raw skills are validated based at least in part on manual review, vetting and validation by human domain experts.

13. The method of claim 9 , further comprising:

extract standardized skills from the set of job postings based on the controlled vocabulary of job skills;

calculate a percent of jobs having each of the job skills across each of a plurality of years; and

generate a smooth trend function based on the percent of jobs having each of the job skills across the plurality of years.

14. A method for classifying job skills, comprising:

identify a set of job postings, a set of job applications, or a text description database;

identify a controlled vocabulary of job skills based on the job postings, the job applications, or the text description database;

generating a synonym file by associating each of the job skill terms with a canonical job skill from the controlled vocabulary of job skills, wherein at least one canonical job skill is associated with a plurality of job skill terms;

extract standardized skills from the job postings, the job applications, or the text description database based on the controlled vocabulary of job skills by identifying skill variants based on the synonym file, wrapping text tokens corresponding to the skill variants in a skill tag, removing non-skill terms, and standardizing the skill variants;

train a machine learning model to extract job skill relationship information including a parent-child relationship between a parent skill and a child skill based on the extracted standardized skills;

create a semantic job skill network based on the machine learning model;

create a job skill taxonomy using the semantic job skill network, wherein the job skill taxonomy comprises a hierarchy of job skills including the parent skill and the child skill of the parent skill;

identify profile information for a job candidate;

identify the child skill as a missing skill for the job candidate based on the profile information and the parent child relationship; and

generate a career trajectory recommendation for the job candidate including a recommendation to obtain the missing skill.

15. The method of claim 14 , the code further comprising instructions executable by the processor to:

identify profile information for a job candidate; and

generate a salary prediction or a career trajectory recommendation for the job candidate based on the profile information and the semantic job skill network.

16. The method of claim 14 , wherein:

the semantic job skill network comprises a plurality of job skills, wherein each of the job skills is associated with a skill name, a skill description, and one or more related skills.

17. The method of claim 14 , the processor being further configured to execute the instructions to:

score each of the set of job postings based on the profile information, the corresponding job description, and the semantic job skill network; and

present one or more of the set of job postings to the job candidate based on the scoring.

18. The method of claim 14 , the processor being further configured to execute the instructions to:

identify a job description for a job posting;

identify a plurality of candidate resumes;

score each of the plurality of candidate resumes based on the job description and the semantic job skill network; and

present one or more of the candidate resumes based on the scoring.