I want to predict salary for each person depending on the job_roles, Technical skills. Column job_roles have 1500 unique roles.

There is a column Technical skills (which have combination of skills) I have applied TFIDF to transform them into numerical(this has created 16000+ columns).

For 1500 unique roles I want to apply hashing method. How do I decide the optimum n_components for unique roles.

  • $\begingroup$ It depends on the cost of a collision. For that reason, it would help us if you could edit your post to include more information about your application. $\endgroup$
    – whuber
    Aug 23 at 13:19
  • $\begingroup$ @whuber I have edited my post. $\endgroup$
    – 10sha25
    Aug 23 at 13:28
  • $\begingroup$ I'm glad you did, because yours is a completely different problem than the original post suggested! The scope of reasonable solutions depends on how much data you have: could you indicate that? $\endgroup$
    – whuber
    Aug 23 at 13:41
  • $\begingroup$ @whuber I am sorry for my previous improper edits. Please read the post now. $\endgroup$
    – 10sha25
    Aug 23 at 14:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.