0
$\begingroup$

This is a book example, but I'm trying to understand it more fully.

So let X $\sim U(0,1)$ and $Z \sim U(0,\frac{1}{10})$. let X,Z be independent uniform distributions. Consider Y=X+Z where the joint distribution $f_{XY}$ can be derived from the joint distribution of (X,Z). So I think this is the joint (X,Z) given $f_{XZ}$ = xz, =(1)(10)=10 for X is uniform, and Z is uniform on $[0,\frac{1}{10}]$

My attempt to work this out is given: Let X=U, Let Y=U+Z so For $f_{XY}(x,y) = f_{XZ}(u,y-u)=u(y-u) = 10$ because y-u is in Z (?). so $f_{XY}=10$ ?

The book explains that by examining the marginal $Y|X=x$ you have Y=x+Z which is uniform on (x,x+$\frac{1}{10})$ and results the $f_{XY}=10$.

my question is how can I show $f_{X+Y}$ using convolution? How I can solve this using convolution is as follows: Let Y=X+Z $f_{X+Y}=\int_{0}^{1} f_{X}(y-t)f_{Z}(t)dt$
$f_{X+Y}=10 \int_{0}^{1} f_{X}(y-t)dt$
where the latter is uniform under [0,1] and is equal to 1 so $f_{X+Y}=10$ not sure if this is correct but reasonable.

any help showing that f$_{XY}=10$ using a convolution approach would be greatly appreciated.

  • 0
    I think it's Casella and Berger, not Burger.2017-02-14
  • 0
    Do you agree that the pdf $f_X(x)=1$ on $[0,1]$ whereas $f_Z(z)=10$ on $[0,1/10]$ (in order that the area under the "curve is 1)? Then...2017-02-14

0 Answers 0