I have a discrete probability distribution: [0.05, 0.05, 0.05, 0.25, 0.6] which represents the probability of a character(given by the index in the vector) to appear.

How can I define a measure between 0 and 1 that shows that most of the probability is concentrated on the last two characters?

  • 2
    $\begingroup$ What is the problem in working with the associated probability measure? $\mathbb P(\textrm{first 3 characters} ) = 0.05 \times 3$ and $\mathbb P(\textrm{last 2 characters} ) = 0.25 + 0.6? $ $\endgroup$ Commented Nov 1, 2022 at 10:56
  • $\begingroup$ I wanted a function which can be applied to different probability distributions $\endgroup$
    – GIONII
    Commented Nov 1, 2022 at 15:23
  • $\begingroup$ The writing of the question didn't apparently say so. Nevertheless, if you could find insight from another post, then all is okay at the end of the day. $\endgroup$ Commented Nov 1, 2022 at 15:32

1 Answer 1


What I was looking for is this: How does one measure the non-uniformity of a distribution?

Entropy is a good measure that differentiates between a uniform distribution (maximum entropy) and a 1 hot encoding (almost 0 entropy).


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.