0
$\begingroup$

Assume that $F_X$ is the CDF of the random variable X and $Q_X$ its quantile function.Prove that $Q_X(F_X(t)) \le t $ and $F_X(Q_X(p)) \ge p $.

I substituted $F_X(t)=P[X \le t]$ and then substituted that in the quantile function $Q_X(p)=min(x:F_X(x) \ge p)$, in this case p is the CDF but i am confused how to go further,seems pretty straightforward but i am missing something. Can some one please help

$\endgroup$
1
$\begingroup$

My first guess is that you might have missed the property that CDF is a non-decreasing function of $x$. Hence: $$ \forall x(x : F_X(x)≥p) \implies F_X(x)≥p $$ It is almost obviously true for a specific $x$ to satisfy the above derivation: $$ x = min(x : F_X(x)≥p) \implies F_X(x)≥p $$ Together with the definition of quantile function, we have: $$ F_X(Q_X(p)) \geqslant p $$ For the first inequality, I would consider a intuitive answer: If $F_X(x)$ is strictly increasing in a neighborhood around point $x = t$, then the equality is attained. The only way to violate the equality is that $x = t$ appears as the right endpoint of a flat region of CDF. In this way, the quantile is the left endpoint of this flat region, which is obviously less than $t$ by their relative position

$\endgroup$

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.