Class Test –STAT 1011Y (11 April 2017)
Instructions:
Answer ALL questions
Time allocated: 35 mins
Silent non-programmable calculators are allowed
Question 1
(a) In a certain insurance company, for a life insurance policy, the policyholder needs to
name two backup beneficiaries. It is known by experience that 10% of policyholders
forget to name two backup beneficiaries. To investigate this issue, he decides to check
each policy, one after another, to see if the policies do not have two backup beneficiaries.
What is the probability that the actuary has to investigate at least seven life insurance
policies to find the first life insurance policy with no two backup beneficiaries? [3]
Let X be the random variable representing the number of life insurance policy investigated
before the first one with no two backup beneficiaries is found.
X~Geo(0.1)
6
P ( X ≥7 )=1−P ( X <7 )=( 1−0.1 ) =0.531
(b) Consider a Poisson random variable X with parameter λ such that
e−λ λ x
f X ( x )= for x=0 , 1 ,2 , …
x!
(i) Show that, for any integer x ≥ 0,
λ
f X ( x+1 )= f (x ) [2]
x +1 X
e−λ λ x+1 e−λ λ x λ λ e− λ λ x λ
f X ( x+1 )= = = = f ( x)
( x +1 ) ! ( x+1 ) x ! (x+ 1) x ! ( x+ 1 ) X
(ii) Show that the mean of X is λ. [5]
∞ −λ x ∞ x ∞ x ∞ x
e λ xλ − λ xλ λ
E ( X ) =∑ x =e−λ ∑ ¿e ∑ =e− λ ∑
x=0 x! x=0 x ! x=1 x ! x=1 x−1 ) !
(
∞ ∞ ∞
−λ λ( y +1) − λ λy λ e− λ λ y
¿e ∑ =e ∑ ¿ λ∑ =λ ×1=λ
y=0 y! y=0 y ! y=0 y!
(iii) A manufacturer has two conveyor belts, one of type C and the other of type D.
The numbers of breakdowns per day on these belts, X and Y, are independent
Poisson random variables with means 1.5 and 0.5 for belts C and D respectively.
Find the conditional probability that if there is exactly one breakdown on a given
day on either belt C or belt D, but not both, then it is conveyor belt C that fails.
[5]
Let X and Y be the [Link] representing the number of breakdowns per day for
conveyor belt C and D respectively.
Class Test –STAT 1011Y (11 April 2017)
X~Po(1.5)
Y~Po(0.5)
P ( 1breakdown onC|Exactly one breakdown ) =?
P ( exactly 1 breakdown )=P ( X=1 , Y =0 )+ P ( X=0 , Y =1 )
¿ P ( X=1 ) × P ( Y =0 )+ P ( X=0 ) × P(Y =1)
¿ 1.5 e−1.5 × e−0.5 +e−1.5 ×0.5 e−0.5¿ 0.2707
P ( X =1, Y =0 )
P ( 1breakdown onC|Exactly one breakdown ) =
P ( exactly 1breakdown )
1.5 e−1.5 × e−0.5
¿ =0.75
0.2707
Question 2
(a) A manufacturer produces bags of sweets. The weights of the bags are independently
distributed with mean 30g and standard deviation 5g. A random sample of 50 bags of
sweets is taken.
(i) State the approximate distribution of the mean weight of the bags in the sample,
quoting any results you use. [2]
Since n is large, the distribution of the mean weight of the bags will approximate to a
normal distribution according to the central limit theorem.
X́ N ( 30,0.5 )
(ii) Use the distribution in part (a) to find the probability that the mean weight of the
bags in the sample is less than 29g. [3]
29−30
(
P ( X́ <29 ) =P Z<
√ 0.5 )
=P ( Z←1.414 )¿ P ( Z >1.414 )=0.0787
(b) The continuous random variable X has probability density function
α −1
f ( x )=α ( 1−x ) , 0< x <1 , α >0
(i) Find the cumulative distribution function, F (x),of X. [3]
x α x
−( 1−x )
F ( x )=∫ α (1−x )
0
α −1
dx=α [ α ]0
=−( 1−x )α −−1¿ 1−( 1−x )
α
0 , x ≤0
{
∴ F ( x )= 1−( 1−x )α , 0< x <1
1 , x ≥1
(ii) Use F (x) to obtain the median of X. [2]
F ( m )=0.5
1 1
∴ 1−( 1−m )α =0.5 ( 1−m)α =0.51−m=0.5 α m=1−0.5 α
Class Test –STAT 1011Y (11 April 2017)
Question 3
(a) State the three desirable properties of a point estimator. [3]
Unbiasedness
Efficiency
Consistency
(b) A candle manufacturer wants to state an average burning time on the packaging of his
perfumed candles. He records the burning times for a random sample of 15 candles and
enters the data into a computer package, which tells him that a 90% confidence interval
for the mean burning time is from 48.388 hours to 56.012 hours.
(i) What are the mean burning time and its standard error for the sample of 15
candles? [4]
We know that the mean is the midpoint of the interval. Therefore,
48.388+56.012
x́= =52.2 hours
2
Therefore we know that x́ +t × s . e=56.012
t=1.7613
56.012−52.2
∴ 52.2+1.7613 × s . e=56.012 s . e= =2.1643 hours
1.7613
(ii) The manufacturer would like to quote a 95% confidence interval for the mean
burning time of his perfumed candles. Calculate this interval and briefly compare
it with the previous interval, stating the assumption made about the burning times
for both to be valid. [3]
For a 95% c.f, t=2.1448
A 95% c.i is given by
¿ ( 47.558,56 .842 ) hours
The interval is wider as we are more confident that this interval contains the true
mean burning time. Both intervals assumed that the burning times come from a
normal population.