0
$\begingroup$

Assuming that I have a list of Users with a list of skills: (each value is a different skill)

And a list of Tasks with a list of demanded skills:

Based on a manual classification that returned: (History list)

  1. Task1 --> Recommended: User3
  2. Task2 --> Reccomended: User5
  3. Task3 --> Recommended: User6
  4. Task4 --> Recommended: User9
  5. Task5 --> Recommended: User8

Given a new Task(6) with skills (A,B,C...):

  • What would be the best way or approach to build a recommendation system that learns with past data (history list) and returns a list of the "best ranked" users to perform this Task?
$\endgroup$
  • $\begingroup$ This is not a recommender system problem. Also in your case what does best mean, what metric are your trying to maximize/minimize ? $\endgroup$ – DollarAkshay Jul 19 '18 at 9:04
1
$\begingroup$

Why do you want to build a learning model?

Your problem seems to be a common assignment problem, all informations are known, absolute and deterministic. You don't need a ML algorithm for this.

Just vectorize your data as one-hot-vectors, define a distance function (you can look to the Hamming's distance).

Compute the distance vector beetween user/task and pick the one who minimize this distance.

| improve this answer | |
$\endgroup$

Your Answer

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

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