Are projections of spherically symmetric distributions always spherically symmetric?

117 Views Asked by At

I originally had this question while I was reading Michael Hardy's high-level proof here: https://math.stackexchange.com/q/56744

He states that "the mapping $(X_1,\ldots, X_n) \mapsto (X_1 - \overline{X}, \ldots, X_n - \overline{X})$ is a projection onto a space of dimension $n-1$. Notice also that its expected value is $0$. Then remember that the probability distribution of the vector $(X_1,\ldots, X_n)$ is spherically symmetric. Therefore so is the distribution of its projection onto a space of dimension one less."

I tried constructing a rigorous proof of why the statement above about spherical symmetry is true, but my linear algebra/geometry chops are not good enough to get me there. Namely, I assume one should work with a projection matrix, but I'm not sure how to tie in the idea of spherical symmetry.

Can someone help me in thinking about how to prove this?

1

There are 1 best solutions below

0
On BEST ANSWER

Any given rotation in an $(n-1)$-dimensional subspace can be lifted to a rotation in the $n$-dimensional space by leaving the coordinate orthogonal to it invariant. Rotating the rotationally symmetric distribution in the $n$-dimensional space by the lifted rotation rotates the projected distribution in the $(n-1)$-dimensional space by the given rotation. Since the symmetric distribution doesn’t change, neither does the projected one.