I am trying to extract entities like university studied at and tech companies from resumes , I have a list of popular universities and companies and I want to find out which university best matches the extracted entity


1) University in the list : IIT Bombay

Extracted entity : Education : Indian Institute of technology Bombay Btech

2)University in the list : Infosys

Extracted entity : Infosys India Ltd.

As you can see , there are extra unwanted words , short forms , expanded forms etc recognized in the extracted entity , is there any sentence similarity algorithms best suited for this purpose ?

Using SpaCy for entity extraction.


1 Answer 1


This is a case of entity resolution for which a standard method is not available. You will have to write your own method also using abbreviation resolution. The python Dedupe package has some distance metrics which you could use to calculate the similarities.

  • $\begingroup$ Please note that it is not just the case of abbreviated names , something like Google LLC and Google India Ltd must have high score in whatever matching algorithm to use $\endgroup$
    – Kitwradr
    Jun 30, 2019 at 18:58
  • $\begingroup$ Yes, that's what exactly entity resolution is about. $\endgroup$ Jul 2, 2019 at 8:31
  • $\begingroup$ So this method is not possible unless I have entity specific datasets? Any way to produce such datasets? $\endgroup$
    – Kitwradr
    Jul 4, 2019 at 5:10
  • $\begingroup$ You don't need entity specific dataset. You can use your current data to act as the dataset. $\endgroup$ Jul 4, 2019 at 16:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.