How to compute the expected value and variance of
\begin{equation} Max(N-B,B) \end{equation}
where $N$ is a constant and $B$ is a binomial variable
$$B\sim B(N,\lambda)$$
also interested in bounds !
How to compute the expected value and variance of
\begin{equation} Max(N-B,B) \end{equation}
where $N$ is a constant and $B$ is a binomial variable
$$B\sim B(N,\lambda)$$
also interested in bounds !
Copyright © 2021 JogjaFile Inc.
Hint: left $A= \max(N-B,B)$, define $X=1$ if $B\ge N/2$, $X=0$ otherwise.
Compute $E[A|X]$ and apply $E[A] = E[E[A|X]]$
Edited:
We have $$ A = \begin{cases} B & \text{if } X=1\\ N-B & \text{if } X=0 \end{cases} $$
Hence we can write $A= X B + (1-X) (N-B)$ and $E[A | X ] = E[B] X + (1-X)(N-E[B])= \lambda X + (1-X) (N-\lambda)$
Then, letting $\mu = E[X] = P(B \ge N/2)$ we have
$$E[A] = E[E[A|X]] = \mu \lambda +(1- \mu)(N-\lambda)$$
To compute $\mu$ you need to evaluate the tail of a Binomial, which has no simple form.