1
$\begingroup$

I have the following question: Let $(X_i)_{i\in\mathbb{N}}$ and $(Y_i)_{i\in\mathbb{N}}$ be real stationary sequences with finite variances and further let $Z$ be a real valued random variable with finite variance, too. We denote by $F_{X_n}(u):=P(X_n\leq u)$ the distribution of $X_n$ and by $\tilde{F}_{Y_n}(u):= \frac{1}{n}\sum_{i=1}^n 1\{Y_i\leq u\}$ the empirical distribution function of $(Y_i)_{i\in\mathbb{N}}$ at $u$. If we know that \begin{align*} X_n\xrightarrow{d}Z \end{align*} ($d$ stands for convergence in distribution) and that \begin{align*} \sup_{u\in\mathbb{R}}|F_{X_n}(u)-\tilde{F}_{Y_n}(u)|\xrightarrow{p}0 \end{align*}($p$ stands for convergence in probability), does it hold then that \begin{align*} \tilde{F}_{Y_n}(u)\xrightarrow{}P(Z\leq u) \end{align*} in some sense (for example in probability) ?

$\endgroup$
1
$\begingroup$

If the distribution of $Z$ is continuous then $\sup_{u\in\mathbb{R}}|F_{X_n}(u)-\tilde{F}_Z(u)|\rightarrow0$ hence $$ \sup_{u\in\mathbb{R}}|\tilde{F}_{Y_n}(u)-P(Z\leqslant u)|\xrightarrow{p}0. $$ Otherwise one might have to delve into the specifics of the setting ($\tilde F_{Y_n}$ random Cesàro means, $(Y_i)$ stationary, etc.).

$\endgroup$

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.