Redirected
edited by
14,589 views
41 votes
41 votes

We are given a set $X = \{X_1,\ldots,X_n\}$ where $X_i=2^i$.  A sample  $S\subseteq X$ is drawn by  selecting each $X_i$  independently with probability $P_i = \frac{1}{2}$ . The expected value of the smallest number in sample $S$ is:

  1. $\left(\frac{1}{n}\right)$
  2. $2$
  3. $\sqrt n$
  4. $n$
edited by

5 Answers

Best answer
108 votes
108 votes
The answer is option $D.$

The smallest element in sample $S$ would be $X_i$  for which $i$ is smallest.

The given probability is for selection of each item of $X$. Independent selection means each item is selected with probability $\frac{1}{2}$.

Probability for $X_1$ to be smallest in $S = \frac{1}{2} $.
Value of $X_1=2$.
Probability for $X_2$ to be smallest in $S$ = Probability of $X_1$ not being in $S$ $\times$ Probability of $X_2$ being in $S$ $= \frac{1}{2} . \frac{1}{2} $.
Value of $X_2=2^2=4$.
Similarly, Probability for $X_i$ to be smallest in $S = (1/2)^i$.

Value of $X_i=2^i$ .

Now Required Expectation=  $\sum_{i=1}^{n}2^{^{i}} \times \left ( \frac{1}{2} \right )^{i} = \sum_{i=1}^n 1 = n $.
edited by
10 votes
10 votes

Ans : D] n

Here, = {2,4,8,...,2n} and S is subset of X.

E[Z] = $\sum$ Z*P[Z] Here, random variable Z is value of smallest number in S, so Z can take value from {2,4,⋯,2n}.

E[smallest number] = $\sum_{i=1}^{n}$ xi * P[xi is smallest]

Now, P[xi is smallest] = $\frac{1}{2^{i}}$ as for xi to be smallest x1, x2,..., xi-1 should not be selected (this has probability of $\frac{1}{2^{i-1}}$ ) and xi should be selected (probability $\frac{1}{2}$ ).

$\therefore$ E[smallest number] = $\sum_{i=1}^{n}$ 2i * $\frac{1}{2^{i}}$ =  $\sum_{i=1}^{n}$ 1 = n

6 votes
6 votes

Events are:

2 is smallest or 4 is smallest or ... or 'n' is smallest number.


Lets consider the probablity that 1 is smallest no. in set S

I choose 2 with probablity 1/2.

Now, I can choose 0 more elements in (n-1)C* $(\frac{1}{2})^{n-1}$

$(\frac{1}{2})^{n-1}$ because all elements were deselected each with probablity 1/2.

OR

I can choose 1 more elements in (n-1)C1 * $(\frac{1}{2})^{n-1}$

(n-1)C1 * $(\frac{1}{2})^{n-1}$ because I selected 1 element with prob. 1/2 and deselected others each with prob. 1/2

OR

.

.

.

I can choose n-1 more elements in (n-1)C1 * $(\frac{1}{2})^{n-1}$

(n-1)C(n-1) * $(\frac{1}{2})^{n-1}$ because I selected (n-1) elements each with prob. 1/2 

Thus, total probablity that 2 is least number

= $\frac{1}{2} * (\frac{1}{2})^{n-1} * (\binom{n-1}{0} + \binom{n-1}{1}+...+\binom{n-1}{n-1})$

=1/2


Similarly, probablity of 4 being least number = 1/4

.

.

.


Similarly, probablity of 2n being least number = $\frac{1}{2^{n}}$

Now, mean

= $\sum x.P(x)$

= 2 * (1/2)  + 4 * (1/4)  + ... + 2n * $\frac{1}{2^{n}}$

= 1 + 1 + 1 + ... + (n times)

= n

3 votes
3 votes
Given, $X = \{2, 4, 8, ……. 2^n\}$ and we have $S \subseteq X$

So we will have different subsets $S$ where $2$ is the smallest , $4$ is the smallest ……. and $2^n$ is the smallest.

Expectation $= 2 * \frac{2^{n-1}}{2^n} + 4 * \frac{2^{n-2}}{2^n} + ……. 2^n * \frac{2^0}{2^n}$

Here,

in $2 * \frac{2^{n-1}}{2^n}$, $\frac{2^{n-1}}{2^n}$ denotes the probability in which $2$ is the smallest number.

$= 1 + 1+ 1+ 1+ 1……. 1$

$= n$
Answer:

Related questions

36 votes
36 votes
7 answers
3
go_editor asked Sep 29, 2014
8,809 views
If the difference between the expectation of the square of a random variable $\left(E\left[X^2\right]\right)$ and the square of the expectation of the random variable $\l...
61 votes
61 votes
3 answers
4
Arjun asked Sep 26, 2014
17,269 views
Suppose you break a stick of unit length at a point chosen uniformly at random. Then the expected length of the shorter stick is ________ .