Suppose $x_1,x_2,...x_n$ are i.i.d. normal random variables, $a_1,a_2,...a_n$ are some positive constants. Could we have following equation $$\sum_{i=1}^n\frac{a_ix_i^2}{\sum_{i=1}^n x_i^2}=O_p(\frac{\sum^n_{i=1}a_i}{n})?$$ How to get this?

  • 1
    $\begingroup$ Hi and welcome! Can you please show some of what you've tried so far? $\endgroup$ – MattAllegro Mar 8 '15 at 10:02

\begin{align} \sum_{i=1}^n\frac{a_ix_i^2}{\sum_{i=1}^n x_i^2} = \frac{(\sum_{i=1}^n a_ix_i^2) /n}{(\sum_{i=1}^n x_i^2)/n} \end{align}

By the strong law of large number, $(\sum_{i=1}^n x_i^2)/n$ converges to 1 almost surely, so we only need to prove $$\frac{(\sum_{i=1}^n a_ix_i^2) /n}{(\sum_{i=1}^n a_i) /n} = O_p(1)$$ i.e.

$$\frac{\sum_{i=1}^n a_ix_i^2}{\sum_{i=1}^n a_i} = O_p(1)$$

Since $E\dfrac{\sum_{i=1}^n a_ix_i^2}{\sum_{i=1}^n a_i} = 1$, we can use $$P(\frac{\sum_{i=1}^n a_ix_i^2}{\sum_{i=1}^n a_i} >M) \leq \frac{1}{M}$$ to conclude


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.