Given $E[f^2(X)] < \infty$ and $X_i \sim_{iid} X$, need to show $$ E\left[ \frac{1}{n} \left( \sum_{i=1}^{n} (f(X_i) - E[f(X)] \right)^2 \right] \leq E[f^2(X)]. $$

My try:

$$E\left[ \frac{1}{n} \left( \sum_{i=1}^{n} (f(X_i) - E[f(X)] \right)^2 \right] = \frac{1}{n}E\left[ \left( \sum_{i=1}^{n} f(X_i) \right)^2 - n E[f(X)] \sum_{i=1}^{n} f(X_i) \right].$$

Now applying Cauchy-Schwarz inequality to the first term we get, $$ E\left[ \frac{1}{n} \left( \sum_{i=1}^{n} (f(X_i) - E[f(X)] \right)^2 \right] \leq n \cdot Var[f(X)] . $$

I am not able to get rid of $n$ on RHS. Any help to fix this is appreciated!


The trick is that your random variables $X_i$ are independent, so that $$E[(f(X_i)-m)(f(X_j)-m)]=E[f(X_i)-m]E[f(X_j)-m]=0,$$ where $m=E[X]$.

This shows that $$E\left[\frac1n\left(\sum_if(X_i)-E[f(X)]\right)^2\right]=\frac1n\sum_iE\left[\left(f(X_i)-E[f(X)]\right)^2\right]=\mathrm{Var}(f(X)).$$


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.