iit−jam mathematical statistics (ms) · pdf fileiit−jam mathematical statistics...
TRANSCRIPT
IITโJAM Mathematical Statistics (MS) 2005
1. Let
๐ท = [
๐ ๐ ๐ ๐ ๐ ๐๐ ๐ ๐ ๐ ๐ ๐๐๐
๐๐
๐ ๐ ๐๐ ๐๐๐ ๐๐ ๐๐ ๐
]
Then the rank of the matrix P is
(a) 1
(b) 2
(c) 3
(d) 4
Solution: (c)
[
1 2 3 4 5 10 โ1 โ2 0 โ4 100
0โ1
0โ2
โ1โ2
0โ4
13
]
๐ 2โฒ = ๐ 2 โ 2๐ 1
๐ 3โฒ = ๐ 3 โ 2๐ 1
๐ 4โฒ = ๐ 4 โ 4๐ 1
~ [
1 2 3 4 5 10 โ1 โ2 0 โ4 100
00
00
โ12
00
12
] ๐ 4โฒ = ๐ 4 โ ๐ 2
~ [
1 2 3 4 5 10 โ1 โ2 0 โ4 100
00
00
โ10
0 10 0
] ๐ 4โฒ = ๐ 4 โ 2๐ 3
๐ ๐๐๐ = 3
2. Consider the following system of linear equations:
๐ + ๐ + ๐ = ๐,
๐ + ๐๐ = ๐,
๐ + ๐๐ = ๐
This system has infinite number of solutions if
(a) ๐ = โ๐, ๐ = ๐
(b) ๐ = ๐, ๐ = ๐
(c) ๐ = ๐, ๐ = ๐
(d) ๐ = โ๐, ๐ = ๐
Solution: (a)
[1 1 11 0 ๐0 1 2
|3๐3] ~ [
1 1 10 โ1 ๐ โ 10 1 2
|3
๐ โ 33
] ๐ 2โฒ = ๐ 2 โ ๐ 1
~ [1 1 10 โ1 ๐ โ 10 1 ๐ + 1
|3
๐ โ 3๐
] ๐ 3โฒ = ๐ 3 + ๐ 2
โ ๐ + 1 = 0; ๐ = 0
โ ๐ = โ1, ๐ = 0
3. Six identical fair dice are thrown independently. Let S denote the number of dice showing
even numbers on their upper faces. Then the variance of the random variable S is
(a) ยฝ
(b) 1
(c) 3/2
(d) 3
Solution: (c) ๐(๐ = ๐ฅ) = (6๐ฅ) (
1
2)๐ฅ(1
2)6โ๐ฅ
; ๐ฅ = 0, 1, 2, 3, 4, 5, 6
โ ๐ ๐๐๐๐๐๐ค๐ ๐๐๐๐๐๐๐๐ ๐๐ ๐ก๐๐๐๐ข๐ก๐๐๐ ๐ค๐๐กโ ๐ = 6, ๐ = 1/2
โ ๐ฃ๐๐๐๐๐๐๐ = ๐๐(1 โ ๐) = 6 ร1
2ร1
2=3
2
4. Let ๐ฟ๐, ๐ฟ๐, โฆ , ๐ฟ๐๐ be a random sample from a distribution having the variance 5. Let
๏ฟฝฬ ๏ฟฝ =๐
๐๐โ๐ฟ๐ ๐๐๐ ๐บ =โ(๐ฟ๐ โ ๏ฟฝฬ ๏ฟฝ)
๐
๐๐
๐=๐
๐๐
๐=๐
Then the value of โ(๐บ) is
(a) 5
(b) 100
(c) 0.25
(d) 105
Solution: (b)
๐ธ [1
๐ โ 1โ(๐ฅ๐ โ ๏ฟฝฬ ๏ฟฝ)
2
๐
๐=1
] = ๐2 โ ๐ธ [1
20โ(๐ฅ๐ โ ๏ฟฝฬ ๏ฟฝ)
2
21
๐=1
] = 5 โ ๐ธ(๐) = 20 ร 5 = 100
5. Let X and Y be independent standard normal random variables. Then the distribution of
๐ผ = (๐ฟ โ ๐
๐ฟ + ๐)๐
๐๐
(a) chiโsquare with 2 degrees of freedom
(b) chiโsquare with 1 degree of freedom
(c) F with (2, 2) degrees of freedom
(d) F with (1, 1) degrees of freedom
Solution: (d)
๐ฅ(1)2
๐ฅ(1)2 = ๐น(1,1)
6. In three independent throws of a fair dice, let X denote the number of upper faces showing
six. Then the value of ๐ฌ(๐ โ ๐ฟ)๐ is
(a) 20/3
(b) 2/3
(c) 5/2
(d) 5/12
Solution: (a)
๐โ๐๐ ๐ = 1, ๐๐๐๐๐๐๐๐๐๐ก๐ฆ ๐๐ (3
1) (1
6) (5
6)2
= 25/72
๐โ๐๐ ๐ = 2, ๐๐๐๐๐๐๐๐๐๐ก๐ฆ ๐๐ (3
2) (1
6)2
(5
6) =
5
72
๐โ๐๐ ๐ = 3, ๐๐๐๐๐๐๐๐๐๐ก๐ฆ ๐๐ (3
3) (1
6)3
=1
216
๐โ๐๐ ๐ = 0, ๐๐๐๐๐๐๐๐๐๐ก๐ฆ ๐๐ (3
0) (5
6)3
=125
216
๐ธ(3 โ ๐)2 =โ(3 โ ๐ฅ)2๐(๐ = ๐ฅ)
=100
72+5
72+ 0 +
125
24=160
24=20
3
7. Let
๐ท = [
๐ ๐ ๐ + ๐ ๐ + ๐๐ ๐ ๐ ๐๐๐
๐ + ๐๐ + ๐
๐๐ + ๐
๐ + ๐๐
]
Then the determinant of the matrix P is
(a) ๐(๐ + ๐)๐
(b) ๐(๐ + ๐)๐
(c) ๐(๐ + ๐)
(d) (๐ + ๐)(๐๐ + ๐)
Solution: (a)
|
1 0 1 + ๐ฅ 1 + ๐ฅ0 1 1 100
1 + ๐ฅ1 + ๐ฅ
โ(1 + ๐ฅ)0
0(1 + ๐ฅ)
| Expanding this w.r.t. 1st column, we will get a 3ร3 matrix, then
perform the given operations:
๐ 3โฒ = ๐ 3 โ ๐ 1 & ๐ 4 โฒ = ๐ 4 โ ๐ 1
det (๐) = |1 1 1
1 + ๐ฅ โ(1 + ๐ฅ) 01 + ๐ฅ 0 โ(1 + ๐ฅ)
| = 3(1 + ๐ฅ)2
8. The area of the region
{(๐, ๐) โถ ๐ โค ๐, ๐ โค ๐,๐
๐โค ๐ + ๐ โค
๐
๐} ๐๐
(a) 9/16
(b) 7/16
(c) 13/32
(d) 19/32
Solution: (d) ๐ ๐๐๐ข๐๐๐๐ ๐๐๐๐ = 1 ร 1 โ1
2ร3
4ร3
4โ1
2ร1
2ร1
2 = 1 โ
9
32=
1
8=
19
32
9. Let E, F and G be three events such that the events E and F are mutually exclusive,
๐ท(๐ฌ โช ๐ญ) = ๐,๐ท(๐ฌ โฉ ๐ฎ) =๐
๐ ๐๐๐ ๐ท(๐ฎ) =
๐
๐๐
Then ๐ท(๐ฌ โช ๐ญ)equals
(a) 1/12
(b) ยผ
(c) 5/12
(d) 1/3
Solution: (d)
๐(๐น โฉ ๐บ) = ๐(๐บ) โ ๐(๐ธ โฉ ๐บ) =7
12โ1
4=
1
3
10. Let X and Y have the joint probability mass function
๐ท(๐ฟ = ๐, ๐ = ๐) =๐
๐๐, ๐ = ๐, ๐,โฆ , ๐; ๐ = ๐, ๐, ๐.
Then the value of the conditional expectation ๐ฌ(๐|๐ฟ = ๐) is
(a) 1
(b) 2
(c) 1.5
(d) 2.5
Solution: (b)
๐ธ(๐|๐ = 3) =โ๐ฆ . ๐(๐ = ๐ฆ;๐ = 3)
๐(๐ = 3)
๐(๐ = ๐ฅ) =1
3๐ฅ(1 + 1 +โฏ+ ๐ฅ ๐ก๐๐๐๐ ) =
1
3
= โ๐ฆ1
3=1 + 2 + 3
3 = 2
3
๐ฆ=1
.
11. Let ๐ฟ๐ ๐๐๐ ๐ฟ๐ be independent random variables with respective moment generating function
๐ด๐(๐) = (๐
๐+๐
๐๐๐)
๐
๐๐๐ ๐ด๐(๐) = ๐๐(๐๐โ๐) , โ โ < ๐ก < โ
Then the value of ๐ท(๐ฟ๐ + ๐ฟ๐ = ๐) is
(a) ๐๐
๐๐๐โ๐
(b) ๐๐
๐๐๐โ๐
(c) ๐๐
๐๐๐โ๐
(d) ๐๐
๐๐๐โ๐
Solution: (a)
๐ธ[๐๐ก๐ฅ] = (3
4+1
4๐๐ก)
3
โ โซ ๐๐ก๐ฅ๐(๐ฅ)๐๐ฅ = (3
4+1
4๐๐ก)
3โ
โโ
๐(๐1 = 0; ๐2 = 1) + ๐(๐1 = 1; ๐2 = 0)
= (3
4)3 21๐โ2
1!+ 3๐ถ1 (
3
4)2
(1
4)๐โ220
0!= (
54
64+27
64) ๐โ2 =
81
64๐โ2
12.
๐ฅ๐ข๐ฆ๐โโ
[๐
๐๐๐๐ช(
๐๐) โซ ๐โ
๐๐๐๐๐โ๐
โ
๐โโ๐๐
๐ ๐] ๐๐๐๐๐๐
(a) 0.5
(b) 0
(c) 0.0228
(d) 0.1587
Solution: (a)
lim๐โโ
โ ๐โ๐ ๐ ๐๐ข๐๐๐ โ ๐๐๐๐๐๐
๐๐๐๐ ๐ก๐ โ = 0.5
๐ โ โ2๐ = ๐ โ 1. Standard deviation.
13. Let ๐ฟ๐ ๐๐๐ ๐ฟ๐ be two independent random variables having the same mean ๐ฝ. Suppose
that ๐ฌ(๐ฟ๐ โ ๐ฝ)๐ = ๐ ๐๐๐ ๐ฌ(๐ฟ๐ โ ๐ฝ)
๐ = ๐. For estimating ๐ฝ, consider the
estimators ๐ป๐ถ(๐ฟ๐, ๐ฟ๐) = ๐ถ๐ฟ๐ โ (๐ โ ๐ถ)๐ฟ๐, ๐ โ [๐, ๐]. The value of ๐ถ โ [๐, ๐],for which the
variance of ๐ป๐ถ(๐ฟ๐, ๐ฟ๐) is minimum, equals
(a) 2/3
(b) ยฝ
(c) ยผ
(d) ยพ
Solution: (a)
๐(๐ผ๐1 โ (1 โ ๐ผ)๐2) = ๐ผ2๐(๐1) + (1 โ ๐ผ)
2๐(๐2) = ๐ผ2 + 2(1 โ ๐ผ)2 = 3๐ผ2 โ 4๐ผ + 2
= 3 [๐ผ2 โ4
3๐ผ +
2
3] = 3 [(๐ผ โ
2
3)2
+2
9]
โ ๐ผ =2
3 ๐๐๐ ๐๐๐๐๐๐ข๐ ๐ฃ๐๐๐๐๐๐๐.
14. Let ๐๐ = ๐, ๐๐ = ๐, ๐๐ = ๐. ๐, ๐๐ = ๐. ๐ be the observed values of a random sample from the
probability density function
๐(๐|๐ฝ) =๐
๐[๐
๐ฝ๐๐๐ฝ +
๐
๐ฝ๐๐โ๐๐ฝ๐ + ๐๐] , ๐ > 0, ๐ > 0
Then the method of moments estimate (MME) of ๐ฝ is
(a) 3.5
(b) 4
(c) 2.5
(d) 1
Solution: (b)
๐ฟ(๐ฅ, ๐) =1
3[1
๐๐๐ฅ๐ +
1
๐2๐โ๐ฅ๐2 + ๐๐ฅ]
๐๐ฟ
๐๐= 0
โโ1
๐2๐๐ฅ/๐ โ
๐ฅ
๐3๐๐ฅ/๐ โ
2
๐3๐โ๐ฅ/๐
2
=โ๐
๐ฅ๐
๐3(๐ฅ + ๐) +
2
๐5๐โ๐ฅ๐2(๐ฅ โ ๐2)
Which is increasing, so ๐ = ๐ฅ๐๐๐ฅ = 4
15. Let ๐๐ = โ๐, ๐๐ = ๐, ๐๐ = ๐, ๐๐ = โ๐ be the observed values of a random sample from the
distribution having the probability density function.
๐(๐|๐ฝ) =๐โ๐
๐๐ฝ โ ๐โ๐ฝ, โ๐ฝ โค ๐ โค ๐ฝ, ๐ฝ > 0.
Then the maximum likelihood estimate of ๐ฝ is
(a) 3
(b) 0.5
(c) 4
(d) Any value between 1 and 2
Solution: (c)
๐ฟ(๐ฅ, ๐) =โ๐โ๐ฅ๐
๐๐ โ ๐โ๐ ; ๐คโ๐๐โ ๐๐ ๐๐๐๐๐๐ข๐ ๐๐ข๐ก ๐๐๐ โ ๐ โค ๐ฅ โค ๐, ๐. ๐. , ๐ = 4
16. Let X and Y be independent and identically distributed uniform random variables over the
interval (0,1) and let ๐บ = ๐ฟ + ๐. Find the probability that the quadratic equation ๐๐๐ + ๐๐บ๐ +
๐ = ๐ has no real root.
Solution:
81๐2 โ 36 < 0 ๐๐๐ฃ๐๐ ๐2 < (6
9)2
โ โ2
3< ๐ <
2
3
๐๐๐๐.=
12 ร
23 ร
23
1 ร 1=2
9
17. Find the number of real roots of the polynomial ๐(๐) = ๐๐ + ๐๐ โ ๐๐ + ๐.
Solution: ๐(๐ฅ) = ๐ฅ5 + ๐ฅ3 โ 2๐ฅ + 1.
Using Descartes rule:
f(x) = + + โ + At most 2 positive.
๐(โ๐ฅ) = โโ + + At most 1 negative
๐โฒ(๐ฅ) = 5๐ฅ4 + 3๐ฅ2 โ 2 = 5 [๐ฅ4 +3
5๐ฅ2 โ
2
5] = 5 [(๐ฅ2 +
3
10)2
โ (7
10)2
]
๐๐ข๐ก๐ก๐๐๐ ๐โฒ(๐ฅ) = 0
โ ๐ฅ2 =4
10=2
5๐ฅ = ยฑโ
2
5
We know complex roots occur in pairs. So, the given equation has 3 real roots.
18. Consider the ๐ ร ๐ matrix
๐ท =
[ ๐
๐ + ๐
๐
๐ + ๐
๐
๐ + ๐โฏ
๐
๐ + ๐๐
๐ + ๐
๐
๐ + ๐
๐
๐ + ๐โฏ
๐
๐ + ๐๐
๐ + ๐โฎ๐
๐ + ๐
๐
๐ + ๐โฎ๐
๐ + ๐
๐
๐ + ๐โฎ๐
๐ + ๐
โฏโฎโฏ
๐
๐ + ๐โฎ๐
๐ + ๐]
Let I be the ๐ ร ๐ identity matrix and let J be the ๐ ร ๐ matrix whose all entries are 1. Express the
matrix P as ๐๐ฐ + ๐๐ฑ, for suitable constants a and b. It is known that the inverse of P is of the form
๐๐ฐ โ ๐ ๐ฑ, for some constants c and d. Find the constants c and d in terms of n.
Solution:
๐ =1
๐ + 1๐ผ +
1
๐ + 1๐ฝ
๐๐โ1 = (๐๐ผ + ๐๐ฝ)(๐๐ผ โ ๐๐ฝ)
๐ผ = ๐๐๐ผ + (๐๐ โ ๐๐)๐ฝ โ ๐๐๐ฝ2 = ๐๐๐ผ + (๐๐ โ ๐๐)๐ฝ โ (๐ + 1)๐๐๐ฝ
โ ๐๐ = 1 & ๐๐ โ ๐๐ โ (๐ + 1)๐๐ = 0
โ ๐ถ = (๐ + 1)
1
๐ + 1(๐ + 1) โ
1
(๐ + 1)๐ โ (๐ + 1)
1
๐ + 1๐ = 0
โ 1 =๐ + 2
๐ + 1๐ โ ๐ =
๐ + 1
๐ + 2
โ ๐โ1 = (๐ + 1)๐ผ โ(๐ + 1)
(๐ + 2)๐ฝ
19. Let ๐(๐) = (๐ + ๐)|๐๐ โ ๐|, โโ < ๐ฅ < โ. Verify the differentiability of the function f(.) at the
points ๐ = โ๐ ๐๐๐ ๐ = ๐.
Solution:
๐(๐ฅ) = (๐ฅ + 1)(๐ฅ2 โ 1) ; ๐ฅ โฅ 1 ๐๐ ๐ฅ โค โ1
= (๐ฅ + 1)(1 โ ๐ฅ2) ; โ1 โค ๐ฅ โค 1
๐โฒ(๐ฅ) = 3๐ฅ2 + 2๐ฅ โ 1 ; ๐ฅ โฅ 1 ๐๐ ๐ฅ โค โ1
= โ3๐ฅ2 โ 2๐ฅ + 1 ; โ1 โค ๐ฅ โค 1
โ ๐(๐ฅ)๐๐ ๐๐๐ก ๐๐๐๐๐๐๐๐๐ก๐๐๐๐๐ ๐๐ก ๐ฅ = โ1 ๐๐๐ ๐ฅ = 1
20. There are four urns labeled ๐ผ๐, ๐ผ๐, ๐ผ๐ ๐๐๐ ๐ผ๐, each containing 3 blue and 5 red balls. The fifth
urn, labeled ๐ผ๐, contains 4 blue and 4 red balls. An urn is selected at random from these five
urns and a ball is drawn at random from it. Given that the selected ball is red, find the probability
that it came from the urn ๐ผ๐.
Solution:
15ร48
15ร48 +
45ร58
=4
24=1
6
21. Let
๐ญ(๐) =
{
๐, ๐๐ ๐ > 0
๐๐
๐๐๐๐ ๐ โค ๐ < 1
๐ + ๐
๐๐(๐๐ โ ๐๐ โ ๐)
๐๐,
๐๐ ๐ โค ๐ < 2๐๐ ๐ โค ๐ โค ๐๐๐ ๐ > 3
Find the value of c for which F(.) is a cumulative distribution function of a random variable X.
Also evaluate ๐ท(๐ โค ๐ฟ โค ๐).
Solution:
๐(๐ฅ) =
{
0 ; ๐ฅ < 0๐ฅ
5; 0 โค ๐ฅ < 1
1
8๐(3 โ ๐ฅ)
0
; 1 โค ๐ฅ < 2; 2 โค ๐ฅ โค 3; ๐ฅ > 3
โ โซ๐ฅ
5๐๐ฅ +โซ
1
8๐๐ฅ + โซ ๐ถ(3 โ ๐ฅ)๐๐ฅ
3
2
2
1
1
0
= 1
โ1
10+1
8+ ๐ถ (3 โ
5
2) = 1
โ๐ถ
2= 1 โ
1
8โ1
10=31
40
โ ๐ถ =31
20
๐(1 โค ๐ฅ < 2) = โซ1
8๐๐ฅ =
1
8
2
1
22. Let A, B and C be pair wise independent events such that ๐ท(๐จ โฉ ๐ฉ) = ๐. ๐ ๐๐๐ ๐ท(๐ฉ โฉ ๐ช) =
๐. ๐. Show that ๐ท(๐จ๐ช โช ๐ช) โฅ ๐/๐.
Solution:
๐(๐ด)๐(๐ต) = 0.1
๐(๐ต)๐(๐ถ) = 0.2
๐(๐ด๐ถ โช ๐ถ) = ๐(๐ด๐ถ) + ๐(๐ถ) โ ๐(๐ด๐ถ)๐(๐ถ)
= 1 โ ๐(๐ด) + ๐(๐ถ) โ ๐(๐ถ)
= 1 โ ๐(๐ด) + ๐(๐ด)๐(๐ถ)
= 1 โ ๐(๐ด)[1 โ ๐(๐ถ)]
= 1 โ ๐(๐ด)[1 โ 2๐(๐ด)]
= 1 โ ๐(๐ด) + 2[๐(๐ด)]2
= 2 [(๐(๐ด))2โ1
2๐(๐ด) +
1
2]
= 2 [{๐(๐ด) โ1
4}2
+7
16]
= 2 [๐(๐ด) โ1
4]2
+7
8 โฅ
7
8
23. Let the random variables X and Y have the joint probability density function
๐(๐, ๐) = {๐๐โ(๐+๐) , ๐ > ๐ฅ > 0
๐, ๐๐๐๐๐๐๐๐๐
Evaluate c and ๐ฌ(๐|๐ฟ = ๐)
Solution:
โซ โซ ๐๐โ(๐ฅ)๐โ๐ฆ๐๐ฅ ๐๐ฆ๐ฆ
0
โ
0
= โซ ๐(1 โ ๐โ๐ฆ)๐โ๐ฆ๐๐ฆ โ
0
= ๐ [โซ ๐โ๐ฆ๐๐ฆ โ โซ ๐โ2๐ฆ๐๐ฆโ
0
โ
0
] = ๐ [1 โ1
2]
๐๐,๐
2= 1
โ ๐ = 2
๐(๐ฅ) = โซ 2๐โ๐ฅ๐โ๐ฆ๐๐ฆ โ
๐ฅ
= 2๐โ2๐ฅ ๐ฅ > 0
๐ธ(๐|๐ = 2) = โซ2๐ฆ๐โ(2+๐ฆ)
2๐โ4๐๐ฆ
โ
2
= โซ ๐ฆ๐โ(๐ฆโ2)๐๐ฆโ
0
= โ๐ฆ๐โ(๐ฆโ2) โ ๐โ(๐ฆโ2) |โ
0= 2 โ 1 = 1
24. Find the area of the region enclosed between the curves
๐ = (๐ โ ๐)๐ ๐๐๐ ๐ = ๐ โ ๐๐
Solution:
(๐ฅ โ 2)2 = 4 โ ๐ฅ2 โ 2๐ฅ2 โ 4๐ฅ = 0 โ 2๐ฅ(๐ฅ โ 2) = 0 โ ๐ฅ = 0, 2
๐ด = โซ (๐ฅ โ 2)2 โ 4 + ๐ฅ2 ๐๐ฅ2
0
=(๐ฅ โ 2)2
3โ 4๐ฅ +
๐ฅ3
3 |2
0 = ๐ด๐๐ ๐ค๐๐
25. Let ๏ฟฝฬ ๏ฟฝ be the sample mean of a random sample of size 10 from a distribution having the
probability density function
๐(๐|๐ฝ) =๐
๐ฝ๐๐๐ฝ , ๐ > 0, ๐ > 0.
Using Chebyshevโs inequality, show that
๐ท(๐ < ๏ฟฝฬ ๏ฟฝ โค ๐๐ฝ) โฅ ๐. ๐
Solution:
๐(๐ฅ|๐) =1
๐๐โ๐ฅ/๐ ; ๐ > 0; ๐ฅ > 0
โ ๐๐ก ๐๐ ๐๐ฅ๐๐๐๐๐๐ก๐๐๐ ๐๐๐ ๐ก๐๐๐๐ข๐ก๐๐๐ ๐ค๐๐กโ ๐๐๐๐๐๐๐ก๐๐1
๐,
โ๐๐๐๐ ๐๐๐๐ ๐ค๐๐๐ ๐๐ 1
1/๐ ๐๐๐, ๐ฃ๐๐๐๐๐๐๐
1
1/๐2 ๐. ๐. , ๐ ๐๐๐ ๐2
So, mean = ๐ ๐๐๐ ๐ . ๐. = ๐
So, by Chebyshevโs inequality
๐(0 < ๏ฟฝฬ ๏ฟฝ < 2๐) > 1 โ1
10= 1 โ 0.1
โ ๐(0 < ๏ฟฝฬ ๏ฟฝ โค 2๐) โฅ 0.9
26. Let X be a continuous random variable having the probability density function
๐(๐) = {๐
๐๐(๐ + ๐) ,โ๐ โค ๐ โค ๐
๐, ๐๐๐๐๐๐๐๐๐
Find the cumulative distribution function ๐ = ๐ฟ๐. Hence derive the expression for probability
density function of Y.
Solution:
Let ๐น(๐ฅ) = ๐(๐ = ๐2 โค ๐ฅ) = ๐(๐2 โค ๐ฅ) = ๐(โโ๐ฅ โค ๐ โค โ๐ฅ)
๐น(๐ฅ) =
{
0 ๐ฅ < 0
โซ2
25(๐ฅ + 2)๐๐ฅ +โซ
2
25(๐ฅ + 2)๐๐ฅ
โ๐ฅ
0
0
โโ๐ฅ
; 0 โค ๐ฅ < 4
โซ2
25(๐ฅ + 2)๐๐ฅ
โ๐ฅ
โ2
1
4 โค ๐ฅ < 9๐ฅ โฅ 9
โน ๐น(๐ฅ) =
{
0; ๐ฅ < 0
8โ๐ฅ
250 โค ๐ฅ < 4
๐ฅ โ 4
25+4(โ๐ฅ + 2)
251;
; 4 โค ๐ฅ < 9๐ฅ โฅ 9
๐. ๐. ๐. ๐๐ ๐ = ๐2
๐(๐ฅ) =
{
0; ๐ฅ โฅ 04
25โ๐ฅ0 โค ๐ฅ โค 4
1
25+
2
25โ๐ฅ0;
; 4 โค ๐ฅ โค 9๐ฅ โฅ 9
27. Let X be a single observation from the probability density function
๐(๐|๐ฝ) = (๐ฝ + ๐)๐๐ฝ , ๐ < ๐ฅ < 1,
Where ๐ฝ โ {๐, ๐} is unknown. Find the most powerful test of level
๐ถ =๐๐
๐๐ ๐๐๐ ๐๐๐๐๐๐๐ ๐ฏ๐ โถ ๐ฝ = ๐ ๐๐๐๐๐๐๐ ๐ฏ๐ โถ ๐ฝ = ๐
What is the power of the test?
Solution: ๐ผ = ๐(๐๐๐๐๐๐ก |๐ป0 ๐๐ ๐ก๐๐ข๐)
๐ป0 ๐๐ ๐ก๐๐ข๐ โ ๐ = 1 โ ๐(๐ฅ|1) = 2๐ฅ, 0 < ๐ฅ < 1
Let critical region be (k, 1), then
โซ ๐(๐ฅ)๐๐ฅ = ๐ผ = โซ 2๐ฅ ๐๐ฅ1
๐
= ๐ผ1
๐
โ 1โ ๐2 =13
49
โ ๐2 =36
49
โ ๐ =6
7
So, acceptance region is (0, 6/7)
๐ฝ = ๐ (๐๐๐๐๐๐ก๐ |๐ป1 ๐๐ ๐ก๐๐ข๐)
๐ป1 ๐๐ ๐ก๐๐ข๐ ๐๐๐ฃ๐๐ ๐ = 2
โ ๐(๐ฅ|2) = 3๐ฅ2; 0 < ๐ฅ < 1
โ ๐ฝ = โซ 3๐ฅ2๐๐ฅ6/7
0
=216
343
โ ๐๐๐ค๐๐ ๐๐ ๐ก๐๐ ๐ก, 1 โ ๐ฝ = 1 โ 216
343 =
133
343 .
28. Let ๐ฟ๐, ๐ฟ๐, โฆ , ๐ฟ๐ be a random sample from a population having the probability density
function
๐(๐|๐ฝ, ๐) = {๐
๐๐โ(๐โ๐ฝ)
๐๐ , ๐ > 0
๐, ๐๐๐๐๐๐๐๐๐
โโ < ๐ < โ, ๐ > 0
Find the method of moments and maximum likelihood estimators of ๐ฝ ๐๐๐ ๐.
Solution: Likelihood function,
๐ฟ(๐ฅ, ๐) =โ1
๐๐โ(๐ฅ๐โ๐)๐2
๐
๐=1
=1
๐๐ ๐โ(โ๐ฅ๐โ๐๐)
๐2
โ log ๐ฟ = โ๐ log๐ โโ๐ฅ๐ + ๐๐
๐2
โ1
๐ฟ.๐๐ฟ
๐๐=๐
๐2> 0 (๐๐๐ค๐๐ฆ๐ )
โ ๐๐๐ฃ๐๐ ๐๐ข๐๐๐ก๐๐๐ ๐๐ ๐๐๐๐๐๐๐ ๐๐๐.
Hence ๐ = ๐ฅ(๐) .
29. Let ๐(๐) = ๐(๐ โ ๐)๐ โ (๐ โ ๐), ๐ โค ๐ โค ๐๐.
Let ๐๐ ๐๐๐ ๐๐ be the points of the global minima and global maxima, respectively, of f(.) in the
interval [0, 20]. Evaluate ๐(๐๐) + ๐(๐๐).
Solution:
๐(๐ฅ) = 3(๐ฅ โ 2)3 โ (๐ฅ โ 2)
โ ๐โฒ(๐ฅ) = 9(๐ฅ โ 2)2 โ 1 = 9๐ฅ2 โ 36๐ฅ + 8
๐๐๐ค ๐โฒ(๐ฅ) = 0
โ 9๐ฅ2 โ 36๐ฅ + 8 = 0
โ ๐ฅ =36 ยฑ โ1296 โ 9 ร 8 ร 4
18
=36 ยฑ 12โ7
18= 2 ยฑ
2โ7
3
๐โฒโฒ(๐ฅ) = 18๐ฅ โ 36 = 18(๐ฅ โ 2)
๐โฒโฒ (2 +2โ7
3) > 0
โ ๐ฅ = 2 +2โ7
3 ๐๐ ๐๐๐๐๐ ๐๐๐๐๐๐ ๐๐๐๐๐ก
๐โฒโฒ (2 โ2โ7
3) < 0
โ ๐ฅ = 2 โ2โ7
3 ๐๐ ๐๐๐๐๐ ๐๐๐ฅ๐๐๐ ๐๐๐๐๐ก.
Clearly ๐ฅ = 0 & ๐ฅ = 20 are global minima & global maxima point (given in the question).
โ ๐ฅ0 = 0 & ๐ฆ0 = 20
๐(๐ฅ0) = ๐(0) = 3(โ2)3 โ (โ2) = โ22
๐(๐ฆ0) = ๐(20) = 3(18)3 โ (18) = 18[3(18)2 โ 1] = 17478
๐๐, ๐(๐ฅ0) + ๐(๐ฆ0) = ๐(0) + ๐(20) = 22 + 17478 = 17456