0
$\begingroup$

Question:

"Let $X_{1},X_{2},\cdots$ be a sequence of independent random variables such that $X_{n}$ is binomial with parameters $2n-1$ and $p=\frac{1}{2}$. If

$$Y_{n} = \frac{2(X_{1}+X_{2}+\cdots +X_{n})}{n}-n$$

show that the sequence $Y_{n}$ converges in distribution to a standard Normal distribution.

My attempt

I believe I'm supposed to use the Strong Law of Large numbers to prove the following:

$$\lim_{n\rightarrow \infty}\frac{2(X_{1}+X_{2}+\cdots +X_{n})}{n}-n=0$$

I've figured that $E[X_{i}]=n-\frac{1}{2}$, but when I try to apply it, I get something like this:

$$2E[X_{i}]=2(n-\frac{1}{2})=2n-1\\ \Rightarrow \lim_{n\rightarrow \infty}(n-1)=\infty$$

Edit: $E[X_{i}]=2i-\frac{1}{2}$ for every $1\leq i \leq n$, which makes the above expectation moot, but I also don't know how to simplify $Y_{n}$ using this expectation. It might be I can't use Law of Strong Number since it means each expectation increases as $n$ increases.

I'm also not sure of how to prove that the variance equals 1.

$\endgroup$
  • 1
    $\begingroup$ You know (are supposed to use) the Central limit theorem? $\endgroup$ – leonbloy Nov 9 '13 at 0:20
  • $\begingroup$ I'll probably need it to completely solve this to prove it's normally distributed. $\endgroup$ – IndustrialPioneer Nov 9 '13 at 0:48
1
$\begingroup$

Forget about the strong law of large numbers. To solve this problem, you need to put the following ideas together.

  • The sum of independent Binomial$(n_i,p)$ random variables is a Binomial$\left(\sum_i n_i,p\right)$ random variable

  • The mean of a sum of random variables is the sum of the means

  • The variance of a sum of independent random variables is the sum of the variances of the random variables

  • If $Z$ is a Binomial$(n,p)$ random variable and $n$ is large, then the cumulative probability distribution function (CDF) of $\displaystyle \left(\frac{Z - E[Z]}{\sqrt{\operatorname{var}(Z)}}\right)$ is well-approximated by (in fact converges to) $\Phi(\cdot)$, the CDF of the standard normal random variable. This is sometimes referred to as the DeMoivre-LaPlace approximation theorem and is a special case of the Central Limit Theorem that leonbloy suggested to you in a comment.

I assume that you do know, or know how to compute, the mean and variance of a binomial random variable.

$\endgroup$

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.