Combined Poisson Random Variable

110 Views Asked by At

A random variable $X$ takes the value $\lambda_A$ with probability $\frac{1}{2}$ and $\lambda_B$ with probability $\frac{1}{2}$. Another random variable $Y$ is ~Poisson($\lambda_A$) if X = $\lambda_A$ and ~Poisson($\lambda_B$) if X = $\lambda_B$.

I need help calculating $Var(Y)$ and $Cov(X, Y)$. I first found $E(Y) = E(Y | X = \lambda_A)P(X = \lambda_A) + E(Y | X = \lambda_B)P(X = \lambda_B)$ = $\frac{1}{2} \lambda_A + \frac{1}{2} \lambda_B$. $E(X) = \frac{1}{2} \lambda_A + \frac{1}{2} \lambda_B$

$E(Y^2) = \frac{1}{2} (\lambda_A^2 + \lambda_A) + \frac{1}{2} (\lambda_B^2 + \lambda_B)$

$E(XY) = E(XY | X = \lambda_A)P(X = \lambda_A) + E(XY | X - \lambda_B)P(X = \lambda_B) = \frac{1}{2} \lambda_A^2 + \frac{1}{2} \lambda_B^2$

So then I can calculate $Var(Y) = E(Y^2) - E(Y)^2$ and $Cov(X, Y) = E(XY) - E(X)E(Y)$?

Am I on the right track for this problem?

1

There are 1 best solutions below

0
On

Generally you are on the right track. Alternatively, you can use Eve's law to compute the variance quickly. Eve's law (or the variance decomposition formula) states:

$$\text{Var}(Y) = E[\text{Var}(Y|X)] + \text{Var}(E[Y|X]).$$

Now, because $Y \sim \text{Poi(X)},$ we have that

$$E[Y|X] = \text{Var(Y|X)} = X.$$

Thus,

$$\text{Var}(Y) = E(X)+\text{Var}(X).$$

Now, there is a similar covariance decomposition formula, but it is just as easy to compute it as you've stated.