The problem I want to solve is as follows: I have data about how many teaching hours different students spend with their teacher in order to pass their exams. The teaching varies from 25-55 hours. In the school there are approximately 15 teachers and given a set of 10 student-teacher features that describe personality characteristics such as, age, sex, education, nationality, etc., I want to develop a student-teacher match model in order to minimize the teaching hours a student needs to pass the exams.

Any idea of what sort of algorithms I should use, or any examples I can look at?

  • $\begingroup$ Maybe for Clarification. Do you envision this as a Classification or as a Minimization Problem: Do you want to find the ideal solution for the current data or do you want to be able to predict it for, lets say, future students? $\endgroup$ – El Burro Jul 1 '19 at 11:58
  • $\begingroup$ Thanks @ElBurro. I think it's a classification problem and yes I want to predict what is the best student-teacher match for incoming future students. $\endgroup$ – Chris Ng Jul 1 '19 at 12:02
  • $\begingroup$ Are there any boundary conditions (student/ teacher ratio- do you want to find the solution that is best for everyone together or rather the solution that is best for one student?) $\endgroup$ – El Burro Jul 1 '19 at 12:05
  • $\begingroup$ I'd like the system to recommend top 3-5 candidate teachers for a new, incoming student. No more conditions. The data I have is pairs of student/teachers and how many teaching hours needed for a specific student with a specific teacher to pass the exams $\endgroup$ – Chris Ng Jul 1 '19 at 12:44
  • 1
    $\begingroup$ How many students do you have in your existing dataset? $\endgroup$ – El Burro Jul 1 '19 at 13:14

This seems to be a recommendation problem. You can find a nice overview here.

  • $\begingroup$ Thanks @ElBurro $\endgroup$ – Chris Ng Jul 1 '19 at 14:08

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.