I'd like to experimentally show that the Bias/Variance tradeoff holds. I tried to do it with modeling a complicated function cos(1/x) using different orders of polynomials but my results were not satisfactory.

Would appreciate any resources to people who's done this experimentally.

enter image description here

  • $\begingroup$ What was unsatisfactory about the results? $\endgroup$ Commented Sep 5, 2016 at 5:18
  • $\begingroup$ Maybe you could check a video from the "Learning from data" MOOC. I think it was covered there $\endgroup$ Commented Sep 7, 2016 at 5:39
  • $\begingroup$ @AlexeyGrigorev seen those -- studied it in Grad school -- but not the formal proof $\endgroup$ Commented Sep 7, 2016 at 21:07


Your Answer

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

Browse other questions tagged or ask your own question.