How to prove the convergence of the series of variances in the theorem to prove strong law of large numbers

70 Views Asked by At

How to prove the convergence of the series of variances theorem in this case?

Let $ \{X_n\}_{n\ge1} $ be a sequence of independent random variables with $ \mathbf{E}X_n = 0 $ for all $ n\ge1 $ and let for some $ p \ge 1 $ \begin{equation} \sum\limits_{n=1}^{\infty} \frac{\mathbf{E}|X_n|^{2p}}{n^{p+1}} < \infty. \end{equation} Prove that $ \frac{S_n}{n} \to 0 $ almost surely as $ n \to \infty $.

I use the next $ \textbf{Theorem} $ with $ b_n = n $ :

Let $ \{X_n\}_{n\ge1} $ be a sequence of independent random variables and let for some $ \varepsilon > 0 $ next conditions hold: \begin{equation} \sum\limits_{n=1}^{\infty} \mathbf{P}(|X_n| \ge \varepsilon b_n) < \infty, \end{equation} \begin{equation} \frac{1}{b_n}\sum\limits_{j=1}^{n} \mathbf{E}Y_j \to 0 \quad \text{as}\quad n \to \infty , \quad Y_j = X_j \mathbf{I}\{ X_j \le \varepsilon b_j \}, \end{equation} \begin{equation} \sum\limits_{n=1}^{\infty} \frac{\mathbf{Var}Y_n}{b_n^2} < \infty. \end{equation} Then $ \frac{S_n}{b_n}\to 0 $ a.s.

I proof the first and the second conditions using exercise's condition. But I can't prove the third. I tried to estimate variance by this way: \begin{equation} \mathbf{Var}Y_n \le \mathbf{E}[X_n^2; \; |X_n| \le n] = \sum\limits_{k=0}^{n-1}\mathbf{E}[X_n^2; \; k < |X_n| \le k+1] \le \sum\limits_{k=0}^{n-1}\frac{1}{k^{2p-2}}\mathbf{E}[X_n^{2p}; \; k < |X_n| \le k+1] \end{equation} But nothing worked.