Cochran’s theorem
Let X be multivariate normally distributed as such that
where each
- 1.
is a quadratic form
- 2.
, where is a by square matrix
- 3.
is positive semidefinite
- 4.
Then any two of the following imply the third:
- 1.
- 2.
each has a chi square distribution (http://planetmath.org/ChiSquaredRandomVariable) with of freedom,
- 3.
’s are mutually independent
As an example, suppose and . Furthermore, assume and , then
This corollary is known as Fisher’s theorem.