Let $x,y\stackrel{\text{i.i.d.}}{\sim}\mathcal{N}(0,I_d)$ be two random $d$-dimensional standard normal distributed vectors and let $\theta$ denote the angle between them. ($I_d$ denotes the $d$-dimensional identity matrix).

Do you know about a probability distribution that models either

  • the distribution of $\theta\in[0,\pi]$, or
  • the distribution of $\cos(\theta)\in[-1,1]$, or
  • the distribution of $\frac{\cos(\theta)+1}{2}\in[0,1]$?

A paper (on page 18, second paragraph) says that the third option is distributed according to $Beta(d/2,d/2)$. Can you derive that? Or can you point me to a resource proving it? Or might it be that they're just using an approximation?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.