exception of order statistics

74 Views Asked by At

Assume $X_1,X_2,\dots ,X_n \overset{i.i.d}{\sim} F(.) $

I define $X_{(1)},X_{(2)},\dots,X_{(n)} $ Like that: $X_{(1)}\leq X_{(2)}\leq \dots\leq X_{(n)}$

For $r\in \{1,2,\dots,n\}$ :

What is the formula of $E[X_{(r)}]$?

I know the distribution function of $X_{(r)}$

$F_r(x)=\sum_{i=r}^{n}\binom{n}{i} (F(x))^i(1-F(x))^{n-i} $

But I still don't know what should I do

i.i.d: independence identical distribution $F_r(x)$: distribution function of $X_{(r)}$

1

There are 1 best solutions below

0
On BEST ANSWER

For continuous independence random variables $X_1,\dots,X_n \sim F(.)$ we can find $f_r(x)$ (density function of $X_{(r)}$):

$f_r(x)=\frac{1}{B(r,n-r+1)}[F(x)]^{r-1}[1-F(x)]^{n-r}f(x)$

Then we can write expectation like that: $E[X_{(r)}]=\int_{-\infty}^{+\infty} xf_r(x) dx$

For discontinuous random variable:

Assume $S^*_X=\{ x:f(x)>0\}$

We can write the $E[X_{(r)}]$ like that:

$E[X_{(r)}]=\sum_{x\in S^*_X}P(X_{(r)}\geq x) =\sum_{x\in S^*_X}1-F_r(x)$

B: Beta function