(Durrett : Probability : Theory and Examples 5th ed, Excercise 3.4.13 ) Lindeberg Feller Theorem

232 Views Asked by At

The problem is

Suppose $P(X_j=j)=P(X_j=-j)=1/2j^{\beta}$ and $P(X_j=0)=1-j^{-\beta}$ where $\beta>0$. Show that if $\beta <1 $ then $S_n/n^{(3-\beta)/2}\Rightarrow c \chi$.

Here is one solution using characteristic function( Let $P(X_j=j)=P(X_j=-j)=1/2j^{\beta}$ and $P(X_j=0)=1-j^{-\beta}$ where $\beta\in(0,1)$, then $S_n/n^{(3-\beta)/2)}\Rightarrow c\chi$ ).

But can we prove this using Lindeberg Feller Theorem?

Letting $X_{n,m} = \frac{X_m}{n^{(3-\beta)/2}}$, it is easy to show that $\lim_{n\rightarrow\infty}\sum\limits^n_{m=1}E[ X_{n,m}^2]\rightarrow \sigma$.

However I can't prove that $$ \lim_{n\rightarrow\infty}\sum\limits^n_{m=1}E[X_{n,m}^2;|X_{n,m}|>\epsilon]\rightarrow 0$$