How to formally calculate the Expected Value of this RV

143 Views Asked by At

Suppose that I have a RV $Z$ that with probability $p$ can be a RV $X$ say $N(\mu,\sigma)$ and with probability $1-p$ can be a RV $Y$ say $\Gamma(\alpha,\beta)$.

Intuitively I think that $E(Z) = pE(X) + (1-p)E(Y)$

but I don't know how to formally show it. Something tells me that it has to do with conditional expectation Could anyone give me some hint on this?

thank you

1

There are 1 best solutions below

4
On BEST ANSWER

This is the law of total expectation, which says that $$ \operatorname{E} (X) = \sum_i{\operatorname{E}(X \mid A_i) \operatorname{P}(A_i)}, $$ where $\{A_i\}_i$ is a finite or countable partition of the sample space (see here for more details).

Suppose that $W$ is a random variable such that $P(W=1)=p$ and $P(W=0)=1-p$. Then $Z=WX+(1-W)Y$ has the distribution that you describe, i.e. $Z$ is equal to $X$ with probability $p$ and to $Y$ with probability $1-p$. Using the law of total expectation, \begin{align*} \operatorname EZ&=\operatorname E(Z\mid W=1)P(W=1)+\operatorname E(Z\mid W=0)P(W=0)\\&=p\operatorname EX+(1-p)\operatorname EY. \end{align*}