Finding the marginal density function of Y

144 Views Asked by At

Okay, the question is like this:

$f_{x}(x) = xe^{-x^2/2}$ for all $x>0$ and $Y = \ln X$, find the density of $Y$.

I don't understand a particular step of this problem.

First they start for $x \longrightarrow x = e^y$ because the natural log is strictly increasing and is therefore one-to-one.

Now this is the step that is confusing me:

$f_y = f_x(v(y))\times|v'(y)|$

where $v(y) = e^y$

Can someone help me grasp the intuition of this step? It is really confusing me. Why is this step used to find the mariginal distribution of Y?

1

There are 1 best solutions below

0
On BEST ANSWER

An easier way see this is by just looking in terms of CDF's.

$F_Y(y) = P(Y \leq y) = P(\log X \leq y) = P(X \leq e^y) = F_X(e^y)$.

Then, $f_Y(y) = F_Y'(y) = \frac{d}{dy} [ F_X(e^y) ] = f_X(e^y) \frac{d}{dy} [ e^y ] = e^y f_X(e^y)$.

That step performs the $F_Y'(y)$ step and accounts for the change of variables you did for writing the CDF of $Y$ (which is an integral of $f_Y(y)$) in terms of the CDF of $X$ (which is also an integral of $f_X(x)$).

If you have a calculus book handy, look up Jacobians - they will talk about them when doing change of variables for integrals.