help with probability density function

44 Views Asked by At

The random variables $X_1, X_2,...$ are independent and $\epsilon \ N(0,1)$. Set

$N = \min\{k \mid X_k \geq 0\}$

Show that the pdf of $X_N$ is

$f_{X_N}(x) = 2\phi(x)$ for $x \geq 0, 0$ elsewhere

Aid: it may be helpful to compute the probability $P(X_N \leq x |N = n)$

The solution is:

$P(X_N \leq x | N = n) = P(X_n \leq x | X_1 \le 0 ,..., X_n \geq 0)$

by independence we get

$ = P(X_n \leq x | X_n \geq 0)$

and as r.v.s are I.I.D

$ = P(X_1 \leq y | X_1 \geq 0)$

and by defintion of conditional probability

$ = \frac{P(0 \leq X_1 \leq y)}{P(X_1 \geq 0)}$

$ = \frac{\phi(x)-\phi(0)}{1/2}$ as $x \geq 0$

$ = 2\cdot (\phi(x) -(\phi(0)) $

Thus $X_N$ is independent of N and we get

$f_{X_N}(x) = \frac{d}{dx}P(X_N \leq x|N = n) = \frac{d}{dx}P(X_N \leq x) = 2\cdot \phi(x)I(x\geq0)$

I wonder the following:

How does $N = \min\{k \mid X_k \geq 0\}$ affect anything ?

And how do we get from

$ = \frac{P(0 \leq X_1 \leq y)}{P(X_1 \geq 0)}$

to

$ = \frac{\phi(x)-\phi(0)}{1/2}$ as $x \geq 0$