Let $\left(X_n\right)_{n\geq 1}$ be i.i.d random variables on $\left(\Omega,\mathcal A, \mathbb P\right)$, $X_1$ with mean $\mu$, and $$ L(\lambda) = \begin{cases} \log\mathbb E\left(e^{\lambda X_1}\right)<\infty, & \text{if }\mathbb E\left(e^{\lambda X_1}\right)<\infty \\ +\infty, & \text{otherwise, } \end{cases} $$

and $\displaystyle L^{*}(x)=\sup\left(x\lambda-L(\lambda)|\lambda\in\mathbb R\right)$

Prove that for any $\alpha > 0$ and $n\geq 1$,

$$\mathbb P\left(\left|\frac{X_1+...+X_n}{n}-\mu\right|\geq \alpha\right)\leq e^{-nL^{*}(\mu+\alpha)}+e^{-nL^{*}(\mu-\alpha)}$$

And deduce the most general law of large numbers you can from the previous inequality.

  • $\begingroup$ Hint: As a first step to deriving the general law of large numbers, you could for example calculate, $L$, $L^{*}$, for $\pm 1$ or Cauchy random variables and see what happens in the inequality. $\endgroup$ – Chris Apr 2 '13 at 23:56
  • $\begingroup$ math.stackexchange.com/questions/349587/… is related. $\endgroup$ – Chris Apr 3 '13 at 1:05

$$\mathbb P\left(\left|\frac{S_n}{n}-\mu\right|\geq\alpha\right)=\mathbb P\left(\frac{S_n}{n}-\mu\geq \alpha\right)+\mathbb P\left(-\frac{S_n}{n}+\mu\geq \alpha\right)$$

So we have to show that:

$$\mathbb P\left(\frac{S_n}{n}-\mu\geq \alpha\right)+\mathbb P\left(-\frac{S_n}{n}+\mu\geq \alpha\right)\leq e^{-nL^{*}(\mu+\alpha)}+e^{-nL^{*}(\mu-\alpha)}$$

$$\mathbb P\left(\frac{S_n}{n}-\mu\geq\alpha\right)\leq e^{-nL^{*}(\mu+\alpha)}$$

Similarly we have

$$\mathbb P\left(-\frac{S_n}{n}+\mu\geq \alpha\right)\leq e^{-nL^{*}(\mu-\alpha)}$$


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.