I am using Weka to make a dataset classification, but there is an option in the classifier evaluation (random seed for XVAL/% split). What does this option mean and what is the seed value? Also, what is the effect of changing the value of this option from one to two or three or other values?

I read that the value of the seed is the starting point, but what is the difference if it is the starting point (seed value) 1, 2, or 10, for example?


1 Answer 1


Seed is just a value by which you can fix the Random Numbers that are being generated in your task. Also, this is a general concept and not just for weka.

So, here random numbers are being used to split the data. Now if you run the code without fixing any seed, you will get different splits on every run. But if you fix the seed to some specific value, you will get the same split every time. This is useful when you want to make your scores reproducable.

  • $\begingroup$ I want to know if the seed value of two is that random values will start from two or not? $\endgroup$ Nov 24, 2019 at 10:09
  • 2
    $\begingroup$ No. Seed value does not represent the start range. $\endgroup$
    – bkshi
    Nov 24, 2019 at 10:12
  • $\begingroup$ So, what is the value of the seed represents in the random generation process ? $\endgroup$ Nov 24, 2019 at 10:25
  • 2
    $\begingroup$ Take a look at the answers here stats.stackexchange.com/questions/354373/… $\endgroup$
    – bkshi
    Nov 24, 2019 at 11:59
  • $\begingroup$ @AhmadSarairah It's a value used to generate the random value. Since random numbers generated from the computer are really pseudo-random, the code that generates them uses the seed as "starting" value. $\endgroup$
    – apaderno
    Jul 17, 2022 at 12:14

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.