Projection-valued measure connection with probability measure

80 Views Asked by At

In wikipedia (https://en.wikipedia.org/wiki/Projection-valued_measure) it is mentioned (in it's notation), that $S_{\pi}(\xi,\xi)$ is a probability measure when $\xi$ has length one. Im wondering why is it true. To check it, I need to check 3 properties of probability measure: 1) probability is non-negative 2) sigma-additivity 3) probability of all space is 1 (which is obvious due to definition of projection-valued measure. So I have a trouble with 1) and 2). (upd. I think the second follows from one of the property of projection-valued measure)

1

There are 1 best solutions below

1
On BEST ANSWER

Note that for any set $E \in M$, we have $S_\pi(\xi, \xi)(E) = \langle \pi(E) \xi, \xi \rangle$. By standard properties of orthogonal projections, $ \langle \pi(E) \xi, \xi \rangle = \langle \pi(E) \xi, \pi(E) \xi \rangle \ge 0$.

As you say, $\sigma$-additivity is part of the definition of $S_\pi$ in Wikipedia's setup.