0

I am studying in class regression analysis and we just went through the classical assumptions. I was wondering, is there a mathematical way to prove that $V(Y_i)=σ^2$ and $Cov(Y_i,Y_j)=0$? Since they are assumptions, maybe there is no way to prove them. However, if a way of proving it existed I would appreciate it very much.

Thanks in advance.

LoneWolf
  • 13
  • 1
    We just assume the $Y_i$ are homoscedastic and uncorrelated. If you look at certain random variables as elements of a vector space, this is (to within a factor $\propto\sigma^2$ dependent on our definition of the inner product) an orthonormality assumption. – J.G. Jun 21 '20 at 12:02
  • What are you asking? There is no 'proof', but maybe some reason behind making these assumptions. – StubbornAtom Jun 21 '20 at 15:08

0 Answers0