Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I am looking at the series $$X_{1,1},$$$$X_{2,1}, X_{2,2}$$ $$X_{3,1},X_{3,2},X_{3,3}$$ $$\dots$$ of independent r.v's with $p_n:=P(X_{n,k}=1)=n^{-\frac{1}{4}}$ and $q_n:=P(X_{n,k}=0)=1-n^{-\frac{1}{4}}$. So they are Bernoulli-distributed.

I would like to know if$$S_n:=\frac{\sum_{k\leq n}(X_{n,k}-E(X_{n,k}))}{Var(\sum_{k\leq n}X_{n,k}) }$$ converges weakly, for $n\rightarrow \infty$.

One can observe that for every $n$ the sums $\sum_{k\leq n}X_{n,k} (=:Y_n)$ are $B(n,p_n)$ distributed. One gets

  • $E(X_{n,k})=p_n$,
  • $E(Y_n)=np_n$,
  • $Var(Y_n)=np_nq_n$.

So it is $$S_n=\frac{Y_n-np_n}{np_nq_n }$$.

The standard CLT can't be applied because the $Y_n$ have different winning-probabilities $p_n$. Also $Y_n$ does not converge to a Poisson-distributed r.v. because $$np_n=n^{\frac{3}{4}}$$ is not constant.

In which way can I apply the CLT?

share|improve this question
    
A very similar question was asked within the last week on MathOverflow. Multiple different approaches were given in the answers. –  cardinal Feb 21 '12 at 18:30
add comment

1 Answer

up vote 2 down vote accepted

I assume your definition of $S_n$ wants a square root in the denominator; otherwise it converges to 0.

You want the Lindeberg-Feller central limit theorem. See Theorem 3.4.5 of R. Durrett, Probability: Theory and Examples (4th edition).

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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