Change of variables and the expected value

65 Views Asked by At

We have a non-negative random variable $z$. We apply a simple location/scale transformation with $a$ and $b$ being the transformation parameters, respectively.

I am trying to show that this change-of-variables argument is true for any function $f$

$$E_{\text{base}}f(a+bz)=E_{\text{transformed}}f(z),$$

where $E$ is the expected value under the base and transformed distributions.

I started with the definition of the expected value $$E_{\text{base}}f(a+bz)=\int f(a+bz)P_{\text{base}}(z)dz$$ and $$E_{\text{transformed}}f(z)=\int f(z)P_{\text{transformed}}(z)dz$$

If we define $u=a+bz$, then we can write $$E_{\text{base}}f(u)=\int f(u)P_{\text{base}}(\frac{u-a}{b})\frac{1}{b}du$$

But now what?

1

There are 1 best solutions below

0
On

Hint: You are done if you can prove, that $$P_{\text{transformed}}(u) = P_{\text{base}}(\frac{u-a}{b})\frac{1}{b}.$$