Suppose that $X\geq0$, and that the moment generating function of $X$ exists in an interval around 0. Given some $\delta>0$ and integer $k=1,2,...$, show that $$\inf_{k=0,1,...}\frac{E(|X|^k)}{\delta^k} \leq \inf_{\lambda>0} \frac{E(e^{\lambda X})}{e^{\lambda \delta}}. $$

Consequently, an optimized bound based on polynomial moments is always at least as good as the Chernoff upper bound. Could anyone enlighten me how to prove this?


Let $b$ denote the LHS. Expanding $e^{\lambda X}$ in a power series you can deduce that $$E(e^{\lambda X}) \ge \sum_{k \ge 0} \frac {b \lambda^k \delta^k}{k!}=b e^{\lambda \delta} \,.$$

  • $\begingroup$ Thank you for letting me have better understanding of this Taylor expansion trick! $\endgroup$
    – Hepdrey
    Jan 12 at 20:21

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.