Scribd
Upload
191 views

Stats Quiz 5 Ans

1. The distribution of the sum of 100 independent uniformly distributed random variables between 0 and 1 is approximately normal with mean 50 due to the central limit theorem. 2. The maximum likelihood estimate of parameter θ in an exponential distribution with density function f(x;θ) = 2x/θ^2, x ≤ θ is the maximum observation in a sample of size 6 which was 0.21. 3. The maximum likelihood estimate of the proportion p of red chips in an urn known to be either 1/5, 1/8, 1/2, or 3/4, given a sample of size 4 containing 1 red and 3 whites, is 1/5.

Uploaded by

Avijit Puri
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
191 views

Stats Quiz 5 Ans

1. The distribution of the sum of 100 independent uniformly distributed random variables between 0 and 1 is approximately normal with mean 50 due to the central limit theorem. 2. The maximum likelihood estimate of parameter θ in an exponential distribution with density function f(x;θ) = 2x/θ^2, x ≤ θ is the maximum observation in a sample of size 6 which was 0.21. 3. The maximum likelihood estimate of the proportion p of red chips in an urn known to be either 1/5, 1/8, 1/2, or 3/4, given a sample of size 4 containing 1 red and 3 whites, is 1/5.

Uploaded by

Avijit Puri
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Answers Quiz 5

Probability and Statistics


Marcia Schafgans
September 2013
1. Consider 100 independent random variables uniformly distributed in the [0, 1[ interval. Call each
of them A
I
. Let 1 =

100
I=1
A
I
. Then, we can say that the distribution of probability of 1 is:
(a) aproximately normal with mean exactly equal to 50.
(b) aproximately normal with mean close (but probably not equal) to 50.
(c) uniform in the [0, 100[ interval.
(d) )
Y
(j) =
_
j,200 for 0 j 0
(100 j),200 for 0 j 100
(e) none of the above.
CLT =)1 Normal
1(1 ) = 1
_

100
I=1
A
I
_
=

100
I=1
1 (A
I
) =

100
I=1
1
2
= 0.
ANSWER: a
2. (LM) What is the MLE for 0 in the jd)
)(j; 0) =
2j
1 0
2
, 0 j 1
if a random sample of size 6 yielded the measurements 0.70, 0.63, 0.92, 0.86, 0.43 and 0.21?
(a) 0
(b) 0.004
(c) 0.18
(d) 0.21
(e) 0.62
1(0) =
_
_
_
6

I=1
2Y
i
10
2
if 0 0.21
0 otherwise
As
6

I=1
2Y
i
10
2
is increasing in 0,

0
1JJ
= 0.21.
ANSWER: d
1
3. (LM) The number of red chips and white chips in an urn is not known, but it is known that
the proportion, j, of reds is either 1, or 1,8 or 1,2 or 8,4. A sample of size 4, drawn with
replacement, yields the sequence red, white, white, white. The MLE for j is:
(a) 1/5
(b) 1/4
(c) 1/3
(d) 1/2
(e) 3/4
1
_
1

_
=
1

_
4

_
3
=
64
62
1
_
1
8
_
=
1
8
_
2
8
_
3
=
8
81
1
_
1
2
_
=
1
2
_
1
2
_
3
=
1
16
1
_
8
4
_
=
8
4
_
1
4
_
3
=
8
26
1(1,) is the biggest among the 4 of them.
Note answer b cannot be the solution (despite
1
4
_
3
4
_
3
exceeding 1(1,)) as it is not in the
sample space... The only feasible solutions of j are 1, or 1,8 or 1,2 or 8,4.
ANSWER: a
4. Suppose a random sample of size : = 6 is drawn from the uniform pdf )(j; 0) = 1,0, 0 j 0 for
the purpose of using

0 = 1
max
to estimate 0. Suppose that 0 = 8.0. Then, Ii(

0 2 [00.8, 00.8[)
is:
(a) 0.40
(b) 0.47
(c) 0.55
(d) 0.60
(e) 0.74
Ii(1
max
< 2, 7) =
6

I=1
Ii(1
I
< 2.7)
=
_
0
10
_
6
= 0.8
So, Ii(1
max
2.7) = 0.47 (and clearly 1
max
is always 8).
ANSWER: b
2
5. (LM) What is the method of moments estimate of 0 if a random sample of size 5 consisting of the
numbers 17, 92, 46, 39 and 56 is drawn from the pdf:
)(j; 0) =
1
0
, 0 j 0
(a) 92
(b) 100
(c) 104
(d) 109
(e) 112
Mean of the distribution: 1(1 ) = 0,2. MME chooses 0 such that
1
n
j
I
= 0,2
Sample mean: j = 0

0
11J
2
= j = 0 =)

0
11J
= 100
ANSWER: b
6. Consider a random variable A (0, o
2
), both 0 and o
2
are unknown. A researcher was trying
to estimate 0 and from a sample of size 0, obtained r = 1 and : = 8. Which one of the following
is a 95% condence interval for 0?
(a) (1, 2.86[
(b) [.01, 7.01[
(c) [4.8, 6.8[
(d) [0.86, 2.86[
(e) [1.81, 1)
r = 1, :,
p
: = 1
t
8,5%
= 1.86, t
8,2.5%
= 2.806
(a) Ii( r 2.86) = Ii(t
8
1.86) = /, correct
ANSWER: a
7. Let 1
1
, 1
2
, ..., 1
n
be a random sample of size : from the jd) )(j; 0) =
1
0
c
/0
, 0 0, j 0.
Consider 2 estimators:

0
1
= 1
1
and

0
2
= :.1
min
.
Hint: the jd) of 1
min
is given by:
)
Y
min
(j) = :.)
Y
(j). [1 1
Y
(j)[
n1
We can say that:
(a) both

0
1
and

0
2
are biased
(b) one of the estimators (

0
1
or

0
2
) is biased, the other is not.
(c) both

0
1
and

0
2
are unbiased and

0
1
is more ecient.
(d) both

0
1
and

0
2
are unbiased and

0
1
and

0
2
are equally ecient.
(e) both

0
1
and

0
2
are unbiased and

0
2
is more ecient.
1
Y
(j) = 1 c
/0
. So:
)
Y
min
(j) = :
1
0
c
/0
.
_
c
/0
_
n1
=
:
0
c


3
which is the pdf of the exponential distribution with mean 0,:.
So,

0
1
: mean is 0, standard deviation is 0.

0
n
: mean is :0,: = 0, standard deviation is :0,: = 0.
ANSWER: d
8. Let A be a normally distributed random variable (A (j, o
2
)). A random sample of 3 real-
izations of A is generated and the sample mean,

A, is calculated. Then, another realization of A
is collected (call it A
4
, it is independent from the 3 previous ones). What is the probability that
A
4
2 [

A 2o,

A 2o[?
(a) 68%
(b) 84%
(c) 91%
(d) 95%
(e) 98%

A (j,
o
2
8
)
So
1(A
4


A) = 0
and as A and

A are independent,
\ ar(A
4


A) = o
2

o
2
8
=
4o
2
8
otd(A
4


A) =
2o
p
8
Now:
Ii(A
4
2 [

A 2o,

A 2o[)
= Ii(A
4


A 2 [2o, 2o[)
= Ii
_
A
4


A
2c
p
3
2
_
2o
2c
p
3
,
2o
2c
p
3
__
= Ii
_
7 2
_

p
8,
p
8
__
where 7 has a standard normal distribution.
As
p
8 = 1.7 and from the table we get 1(1.7) = 0.04, 1 1(1.7) = 0.044
`
6
So, Ii
_
7 2
_

p
8,
p
8
_
= 0.04 0.044
`
6 = 0.0108
ANSWER: c
4

You might also like