Sorry if my question is simple, I have a data set with two class and want to check and set normal distribution on it(if it was necessary, in MATLAB).

But the question is that I should use it on every rows or every column, it's 200 rows and 30 columns.


1 Answer 1


It's hard to be sure of anything without having a glimpse at your data but what I assume from what you are saying is:

  • you have 2 classes which probably are given as 1 columns

  • you have 29 features

  • you have 200 samples

    Edit: Then while writing my answer I'm not too sure what you wanna do. If you have 2 classes it is likely that you are about to do some classification tasks so I don't really get what you want to modelize with a normal distribution.

  • $\begingroup$ Thanks for you interesting and time, I'm working on a classification model to get a good accuracy to classify my data with that 2 classes $\endgroup$
    – Davood
    Commented Dec 2, 2019 at 11:18
  • $\begingroup$ So you what you want to check is if your features follow a normal distribution? what do you mean by "set normal distribution on it"? $\endgroup$
    – Jeanba
    Commented Dec 2, 2019 at 11:27
  • $\begingroup$ I was reading some content on the Internet to resolve the over fitting on my algorithm and It seems when your data set has not a normal distribution it can cause over fitting because of it I wanted to check and resolve it $\endgroup$
    – Davood
    Commented Dec 2, 2019 at 11:46
  • $\begingroup$ "I was reading some content", I think it would be helpful to provide some source here, as such I am clueless on what you are intending to do. Or maybe are you trying to check if your samples are evenly distributed among your two classes? But then there is no normal distribution involved. $\endgroup$
    – Jeanba
    Commented Dec 2, 2019 at 11:58
  • $\begingroup$ Btw 200 rows is really few data; What make you think you are over-fitting? What kind of models have you implemented/tried to implement so far? $\endgroup$
    – Jeanba
    Commented Dec 2, 2019 at 12:02

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.