Distribution of two independent standard normals

88 Views Asked by At

Suppose that $X$ and $Y$ are distributed as independent Standard Normals. Find the distribution of $(X-Y, X+Y)$.

Isn't the case for $X-Y$ elementary? Since they are both standard normals, this should equate to $0$ for all components concerning this portion of the vector?

Additionally, when it says "find the distribution", would this just mean me finding the mean, variance and covariance of each component?

1

There are 1 best solutions below

0
On BEST ANSWER

If $X, Y$ are independent and have the standard normal distribution, then $X-Y$ has normal distribution of mean $E(X-Y)=E(X)-E(Y)=0$ and dispersion $\sigma^2(X-Y)= \sigma^2(X)+\sigma^2(Y)=2$. Analogously, $X+Y\sim N(E(X)+E(Y), \sigma^2(X)+\sigma^2(Y))$, i.e. $X+Y\sim N(0,\sigma^2=2)$.

In order to get the joint distribution of $(X-Y, X+Y)$ we have to compute its mean vector and the matrix of covariance, $\Sigma$. The mean vector $m=(0,0)^T$. To get $\Sigma_{12}=\Sigma_{21}$ we need the covariance: $cov(X-Y, X+Y)=E((X-Y)(X+Y))-E(X-Y)E(X+Y)=E(X^2-Y^2)-[(E(X))^2-(E(Y))^2]=E(X^2)-(E(X))^2-[E(Y^2)-(E(Y))^2]=\sigma^2(X)-\sigma^2(Y)=0$. Hence the two random variables are uncorelated, and thus $\Sigma=\left[\begin{array}{cc} \sigma^2(X-Y)&0\\ 0&\sigma^2(X+Y)\end{array}\right]=\left[\begin{array}{cc}2&0\\0&2\end{array}\right]$. Now you can write down the joint density distribution function.