2

I am a developer and having little knowledge in text parsing. I need to parse the Job description and get some outputs. I need to parse the following fields from Job description.

Job Responsibilities, Qualification, Specialization, Domain, Skills Required, Job Description, Work Experience Min, Work Experience Max, Industry, Occupation, Functional Area, Currency, Salary, Salary Type, Employment Type, Work Authorisation, Required Visa Status, Required English Level, Country, State, City, Zipcode, Address of Job.

To accomplish this, I am utilizing the Regex pattern matching. But the output efficiency is low many times. It sometimes requires exact pattern to identify the parameters. So it fails many times.

I found other ways too.

  1. Named Entity Recognition: By using Stanford NLp, I am able to get the location, address. But I don't know how can I train the module for other parameters or we have any possibilities.
  2. Fuzzy logic: Did some research on fuzzy logic to validate the results.

My questions are,

1. What are the approaches to accomplish the JD parsing? 

2. How effective is NER? 

3. Is there any conceivable outcomes to use fuzzy logic in JD text parsing?

Any help would be really appriciated.

Raj De Inno
  • 1,092
  • 2
  • 14
  • 33

1 Answers1

0

I think you can try dependency parsing if regex doesn't work accurately. NER will not support all the findings you need. Employment type is something would like to learn from you as well.

Prateek
  • 1