$L_2$ norm of the generalized inverse of the empirical cumulative distribution function

61 Views Asked by At

Can anyone give me some insight on how to solve this problem? Any help would be greatly appreciated !

Let $(X_n)_{n \geq 1}$ an i.i.d. sequence of real valued r.v.'s with c.d.f. $F$ which is stricly increasing and has two derivatives everywhere. Let $F_n(x) = \frac{1}{n}\sum_{i=1}^n {\bf 1}_{X_i \leq x}$ be the empirical c.d.f., $F_n^{(-1)}(u) = \inf\{x \in \mathbb{R} | F_n(x) \geq u\}$ be the generalized inverse of the empirical cumulative distribution function and $F^{-1}$ be the inverse of $F$. Note that $F_n^{(-1)}(u) = X_{(k_n)}$ where $k_n = \lceil nu \rceil$.

Is this true that under the above hypothesis, \begin{equation} \mathbb{E}\big[\big(F_n^{(-1)}(u) - F^{-1}(u)\big)^2\big] \end{equation} is uniformly bounded?

Thank you