I have time domain data which is having binary label in form of 0 and 1. I applied FFT to all the features and label as well. The problem is my output label is now not binary anymore. How to solve this classification problem?


1 Answer 1


If you are dealing with a classification problem, you should treat the $0$ and $1$ as labels.

The goal of FFT is to convert time domain to frequency domain.

You do not want to include the label as part of your FFT procedure.

  • $\begingroup$ Thanks for your response. I figured it out that i need to apply fft on a vector of values of same class. Is this the right procedure? $\endgroup$ Commented Apr 29, 2019 at 4:35
  • $\begingroup$ can't tell for sure if something is the right procedure in data science as it is a broad area, give it a try and see the outcome? $\endgroup$ Commented Apr 29, 2019 at 9:40
  • $\begingroup$ Yeah i did and now i am preserving some information. Thanks for your inputs. $\endgroup$ Commented Apr 29, 2019 at 9:45

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.