0
$\begingroup$

Let $X_i$ be a sequence of independent and identically distributed random variables with $\mathbb{E}(X_i) = \mu$ and var($X_i$) $= 1$ for all $i$. Define $S_n = \frac{1}{n}\sum_{i-1}^n X_i$ and

$$ g(x) = \begin{cases} x & \text{if } x < 1 \\ 2x-1 & \text{if } x \geq 1 \end{cases} $$

Does $\sqrt{n}(g(S_n)-g(\mu))$ converge to anything when $\mu=1$?

I know that for $\mu \neq 1$, I can apply the delta method and obtain convergence in distribution. When $\mu = 1$ it seems like it converges to a "piecewise" distribution. Is there anything to this?

$\endgroup$
0
$\begingroup$

Let $\{X_n\}$ be i.i.d. positive random variables with mean 1 and variance 1. (For example, they can have exponential distribution with parameter 1). By strong law $S_N \to \infty$ almost surely. This implies $g(S_n) \to \infty$ almost surely and $\sqrt {n} (g(S_n)-1) \to \infty$ almost surely.

$\endgroup$
  • $\begingroup$ Sorry I made a mistake. S_n should be the average, not simply a sum. $\endgroup$ – is it normal Feb 10 '18 at 17:06

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.