4
$\begingroup$

Say we have a binomial random variable $B \sim \operatorname{Bin}(n,p)$ and $n$ i.i.d. Normal random variables $N_i \sim N(p, 100)$. What is the distribution of $\sum_{i=1}^B N_i$?

I'm not really sure how to go about this. I know that I can calculate the expectation through Wald's Lemma, but in general, I'm not sure how to deal with random sums like this.

$\endgroup$
7
  • 1
    $\begingroup$ What have you tried? Do you have a general strategy for approaching problems like this? Is this a homework problem? If so, you should add the self-study tag and read the corresponding wiki $\endgroup$
    – jld
    Oct 25 '17 at 0:57
  • $\begingroup$ Thank you for the feedback. I edited my response to answer these questions. Where can I find the corresponding wiki? $\endgroup$
    – Robin
    Oct 25 '17 at 1:00
  • 1
    $\begingroup$ thanks for the edits. You can just click on self-study and then click on the learn more $\endgroup$
    – jld
    Oct 25 '17 at 1:02
  • $\begingroup$ Let $Z = \sum_{i=1}^B N_i$. You want $P(Z \leq z)$ for $z \in \mathbb{R}$. Expand $P(Z)$ using the tower law, then you have many different sums of iid normal RVs which are still normal. $\endgroup$
    – Alex
    Oct 25 '17 at 1:03
  • $\begingroup$ The random variable $\sum_{i=1}^{B}N_{i}$ is called a compound random variable and its mean is $E(B)E(N)$ and variance $E[B]\cdot Var[N] + Var[B]\cdot \left\{E[N]\right\}^{2}$. $\endgroup$
    – L.V.Rao
    Oct 25 '17 at 1:03
4
$\begingroup$

I'm not sure how to deal with random sums like this

Let $X_{1}, X_{2}, X_{3},\cdots$ be independent and identically distributed random variables and let $N$ be a non-negative integer valued random variable and is independent of the random variable $X$. If $$S_{N}=X_{1}+X_{2}+\cdots +X_{N}$$ be the sum of a random number of independent and identically distributed random variables, then \begin{eqnarray*} % \nonumber to remove numbering (before each equation) E[S_{N}]&=&E[N]\cdot E[X]\\ Var[S_{N}]&=& E[N]\cdot Var[X] + Var[N]\cdot \left\{E[X]\right\}^{2} \end{eqnarray*} We first find the conditional expectation of $S_{N}$ given the value of $N$ as \begin{eqnarray*} % \nonumber to remove numbering (before each equation) E\left[S_{N}|N=n\right] &=& E\left[X_{1}+X_{2}+\cdots +X_{N}|N=n\right] \\ &=& E\left[X_{1}+X_{2}+\cdots +X_{n}\right]\\ & &\quad \mbox{ since X and N are independent RV's}\\ &=& n E[X], \quad\quad\mbox{ since $X_{i}'s$ are iid RV's}\\ E\left[S_{N}\right] &=& \sum_{n}\underbrace{E\left(S_{N}|N=n\right)}_{n\cdot E[X]} P\left\{N=n\right\}\\ &=& \sum_{n} n\cdot E[X]\cdot P\left\{N=n\right\}\\ &=&E(X)\cdot \sum_{n}n\cdot P\left\{N=n\right\}\\ E\left[S_{N}\right] &=& E[X]\cdot E[N] \end{eqnarray*} In order to find an expression for $Var[S_{N}]$, we first find the second conditional moment of $S_{N}$. \begin{eqnarray*} % \nonumber to remove numbering (before each equation) E(S_{N}^{2}|N=n) &=& E[(X_{1}+X_{2}+\cdots +X_{N})^{2}|N=n] \\ &=& E[(X_{1}+X_{2}+\cdots +X_{n})^{2}]\\ &=& E\left\{\sum_{i=1}^{n}X_{i}^{2}+2\sum_{i<j}X_{i}X_{j}\right\}\\ &=& \sum_{i=1}^{n}E\left(X_{i}^{2}\right) + 2 \sum_{i<j} E\left(X_{i}\right)E\left(X_{j}\right)\\ &=& n\cdot E\left(X^{2}\right) + 2\cdot \frac{n(n-1)}{2}E\left(X\right)E\left(X\right)\\ \end{eqnarray*} \begin{eqnarray*} E(S_{N}^{2}|N=n) &=& n\cdot E\left(X^{2}\right) + n(n-1)\left(E(X)\right)^{2}\\ &=& n[E(X^{2})-(E(X))^{2}] + n^{2}(E(X))^{2}\\ E(S_{N}^{2}|N=n) &=& n Var(X) + n^{2}(E(X))^{2} \end{eqnarray*} Now \begin{eqnarray*} % \nonumber to remove numbering (before each equation) Var(S_{N}) &=& E(S_{N}^{2}) - \left[E(S_{N})\right]^{2} \\ &=& \sum_{n}E(S_{N}^{2}|N=n)P\{N=n\}-[E(X)E(N)]^{2}\\ &=&\sum_{n}\left[nVar\left(X\right)+n^{2}\left(E\left(X\right)\right)^{2}\right] P\left\{N=n\right\}\\ & &\quad\quad\quad -\left[E(X)E(N)\right]^{2}\\ &=& Var(X) \sum_{n}nP\left\{N=n\right\}\\ & & + \left(E\left(X\right)\right)^{2}\sum_{n}n^{2}P\left\{N=n\right\} - \left[E(X)E(N)\right]^{2}\\ &=&E(N)Var(X) + \left(E\left(X\right)\right)^{2} \left[E(N^{2})-(E(N))^{2}\right]\\ Var(S_{N}) &=& E(N)Var(X) + Var[N](E(X))^{2} \end{eqnarray*}

$\endgroup$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.