CH 5 [central limit theorem]

Ans1. X represents no. of claims arising in home insurance policy

then X follow poisson ( 0.075 )

so , p ¿

¿ by normal approximation
500

∑ X i follow normal (500 ( 0.075 ) ,¿500 (0.075))¿

x=1

¿, p¿

p ( Normal ( 37.5 ,37.5 )> 49.5 ) = p Z>

( 49.5−37.5
√37.5 )
=1− p (Z ≤1.959)

using actuarial table 1− p ( Z ≤ 1.959 )=1−0.974941

¿ 0.025059

Ans2. X follow gamma ( 10 ,0.2 )

( a ) p ( X > 80 ) [ using normal distribution ]

by normal approximation X follow Normal

( 10 10
,
0.2 0.22 )
so , X follow Normal (50,250 )

then , p ( X > 80 )= p ( Normal ( 50,250 ) >80 )= p Z >

( √ 250 ))
( 80−50

¿ 1− p ( Z ≤ 1.897 )=1−0.971082

¿ 0.028918

(b) p ( X >80 ) [ using chi square distribution ]

if X follow gamma ( 10 , 0.2 ) then by actuarialtable page 12
2 2
2 ( 0.2 ) X follow χ 2 (10)∨, 0.4 X follow χ 20

so , p ( X >80 ) =p ( χ 220> 0.4 ( 80 )) = p ( χ 220>32 )

¿ 1−0.9567=0.0433

As, ¿ part ( a ) , weuse approximation∧¿ part ( b ) we just

transform actual distributionhence , answer ∈ part ( b )
 i .e , 0.0433is more accurate

Ans3. continuity correction

( i ) X <6.5

( ii ) −0.5< X < 0.5

( iii ) X >−2.5

( iv ) 5.5< X< 10.5

( v ) 2.5< X <7.5

( vi ) 4 ≤10 X< 48

0.4 ≤ X < 4.8

as X can only take integer value so , 1≤ X <5

¿ continuity correction of 1≤ X < 4 is 0.5< X <4.5

Ans4. X represents no . of policies elapse before they expires

X follow bin ( n=100 , p=0.15 )

then by normal approximation X follow normal ¿

¿ , p ( X >20 ) =p ( X >20.5 ) [ by continuity correction ]

(
hence , p ( X > 20.5 )= p Z >
( 20.5−( 100 ( 0.15 ) ) )
√ 100 ( 0.15 ) ( 0.85 ) )
¿ p ( Z >1.54 )

¿ 1− p ( Z< 1.54 )

¿ 1−0.93822

¿ 0.06178

so , approximate probability that more than 20 policies elapse

before they expires is0.062

Ans5. N represents no . of clients who do not find brochure helpful

N follow bin ( n=1000 , p=0.05 )

then by normal approximation N follow Normal ( 1000 ( 0.05 ) , 1000 ( 0.05 ) ( 0.95 ) )
 ¿ , N follow Normal (50 , 47.5)

¿ by continuity correction p ( 40< N <70 ) =p ( 40.5< N <69.5 )

hence , p ( 40.5< N <69.5 ) =p

( 40.5−50
√ 47.5
<Z<
√47.5 )
69.5−50

¿ p (−1.378< Z< 2.83 )

¿ 0.99767−( 1−0.9159 )=0.91357

so , approximate probability that ∈sample of 1000 no . of clients

who do not find brochure helpful , lies between 40 ¿ 70 is0.91357

Ans6.let X represents no . of people having blood group A

X follow binomial ( n=300 , p=0.45 )

then by normal approximation

X follow Normal ( 300 ( 0.45 ) , 300 ( 0.45 ) ( 0.55 ) )∨N (135 ,74.25)

¿ by continuity correction p ( X> 115 ) =p ( X >115.5 )

(
So, p ( X >115.5 )= p Z >
115.5−135
√ 74.25 )
¿ ( Z>−2.263 )

¿ 0.988183

so , 0.988183 isthe probability that ∈sample of 300 , more than

115 have blood group A

Ans7.let X follow Normal ( 25 , 4 )

16 samples are drawn ¿ X

then ,by central limit theorem , X follow Normal 25 ,

( 16 ))
( 4

Or, X follow Normal ( 25 , 0.25 )

(
then , p ( X >26 )= p Z >
26−25
√0.25 )
= p ( Z >2 )=0.02275

Ans8. X represents claim amount assured under each policy

2
X has mean=8000∧variance=3000
 100
¿ by central limit theorem ∑ X i follow Normal ( 100 ( 8000 ) , 100 ( 3000 ) )
2

i=1

100

or∑ X i follow Normal ( 800000,100 (3000 ) )

2

i=1

so , p ¿

¿ p ( Z >1.5 )

¿ 1−0.93319=0.06681

Ans9. let each selectionbe a random variable X

1
where , X takes 1 ,2 , 3 , 4∧5 with probability of each
5

then E ( X )=( 1+2+3+ 4+5 )∗ ( 15 )=3

¿ E ( X ) =( 1 +2 +3 +4 + 5 )∗( )=11
2 2 2 2 2 1 2
5
2
so , var ( X )=11− ( 3 ) =2

now , by central limit theorem

100

∑ X i follow Normal ( ( 100 ) 3 , ( 100 ) 2 )

i=1

( ) ( )
100 100
so , p 280 ≤ ∑ X i ≤320 = p 279.5< ∑ X i< 320.5
i=1 i=1

¿p
( 279.5−300
√ 200
<Z<
320.5−300
√200 )
= p (−1.449<Z <1.449 )

¿ 2 ( 0.92633 )−1=0.85266

so , probability that ∑ of 100 sample lies between 280¿ 320

Is approximately 0.85266
Ans10. Y follow gamma ( α , λ ) where , α >1
α
λ α −1 −λy
thenusing table , f ( y )= y e ; y >0
(
gamma α )

( i ) ( a ) ¿ find that for which value of y , f ( y ) is maximum

take log on both sides ,

 ln ( f ( y ) )=ln ( c ) + ( α−1 ) ln ( y ) −λy [where c is some constant ]

differentiating both sides w .r . t y

d ln ( f ( y ) ) ( α −1 )
=0+ −λ
dy y
for value of y

( α −1 ) α−1
−λ=0 gives y =
y λ
( α −1 )
¿ its double derivative i . e ,− 2
<0
y
α −1
y= is value for which f ( y ) is maximum
y

(i) (b) As , gamma distribution is positively skewed distribution

so , its mode>mean

α−1 α 1
¿ mode= = −
λ λ λ
α
where , is mean μ
λ
1 μ
so , mode=μ− =μ− but as α tends ¿ ∞ then,
λ α

[
mode=μ because ,
μ
α
tends ¿ 0
]
so , we can say our gamma ( α , λ ) tends¿ normal μ , ( μλ )
(ii) if Y follow gamma (50 ,0.2 )
( a ) then usingchi sqaure
2
we can say 0.4 Y follow χ 100

so , p ( Y > 350 )= p ( χ 100

2
>0.4 (350 ) )

¿ p ( χ 2100 >140 )

(
¿ , p χ 225 >
140
4 )= p ( χ 225> 35 ) =1−0.9118=0.0882

( b ) by central limit theorem

 Y follow Normal
( 50 50
,
0.2 0.22 )
¿ , Y follow Normal ( 250,1250 )

so , p ( Y > 350 )= p Z >

( 350−250
√1250 )
=1− p ( Z< 2.828 )

=0.002344
(iii)difference between answers ∈ ( ii ) ( a ) ∧( ii ) ( b )
is because we use approximation∈ ( ii ) ( b )

¿ there is no approximation∈ part ( ii ) ( a )

hence , we can see that difference .

Ans11. X i follow gamma ( α , λ )

)=E ( e
( i ) M X ( t )=E ( e tx ( )) t
∑ xi
n

¿ E (e )
t
( X + X + X … … X n)
n 1 2 3

[ ( )]
n
t
(X )
n 1
¿ E e

[ ( )]
n
t
¿ MX i
n

( ) t < λ [ ¿ actuarial table ]

−α
t
as M x ( t )= 1−
λ

( nt )=(1− nλt )
−α
so , M X i

[( ) ] ( )
−α n −n α
t t
hence , M X ( t )= 1− = 1−
nλ nλ

then ,by uniqueness property ,

we can say , X follow gamma ( nα , nλ )

( 1
)
(ii) T follow exp 2000 hours then T also follow gamma 1 , 2000 ( ( )) 1

so , usingresult of previous part we can say

 (
T follow gamma 10 ( 1 ) , 10 ( 2000
1
))
thenusing chi sqaure conversion

2 ( 2001 ) T follow χ 2
20

( ( ) ( ))
so , p ( T > 4000 )= p χ 20>
2 1
100
4000

¿ p ( χ 220 > 40 )

¿ 1−0.9950=0.005

so , probability that average time of 10 bulbs exceed 4000 hours

is 0.5 %