FACULTY OF SCIENCE AND ENGINEERING

DEPARTMENT OF MATHEMATICS AND STATISTICS

END OF SEMESTER ASSESSMENT PAPER

MODULE CODE: MA4004 SEMESTER: Spring 2018

MODULE TITLE: Engineering Mathematics 4 DURATION OF EXAM: 2.5 hours

LECTURER: Dr. Kevin Burke GRADING SCHEME: 70% of module

EXTERNAL EXAMINER: Prof. Adele Marshall

INSTRUCTIONS TO CANDIDATE

• This exam consists of two sections: (I) Probability and (II) Statistics.

• Attempt two questions from each section, i.e., four questions in total.
All questions carry equal marks.

• All work must be shown clearly and logically using appropriate symbols and
probability notation. Failure to do so will lose marks.

• Write down the formula you intend to use at each stage before filling it in
with numbers.

• Formula sheets are provided at the back of this exam paper.

• Statistical tables are available from the invigilators.

Part I: Probability Question 1

a) Assume that A and B are independent events with Pr(A) = 0.65 and
Pr(B) = 0.88. Furthermore, let C be an event which is necessary for A,
i.e., given that A has occurred, C is certain to have occurred.
Calculate the following:
i) Pr(A ∪ B) ; (2 marks)

ii) Pr(Ac ∪ B c ) ; (2 marks)

iii) Pr(Ac | B) ; (2 marks)

iv) Pr(A ∩ C) ; (2 marks)

v) Pr(C) when Pr(A | C) = 0.8 . (2 marks)

b) Michelle inspects devices produced on a factory line. Assume that 20%

of these devices are faulty. Furthermore, if a device is faulty, Michelle
classifies it as being faulty 90% of the time, whereas, if a device is not
faulty, she classifies it as being faulty 5% of the time.
i) Define the following events
• F = “the device is faulty”, and
• C = “Michelle classifies the device as being faulty”.
and, hence, convert the above information into probability statements.
(3 marks)

ii) What proportion of faulty devices does Michelle miss in her inspection?
Note: these are the faulty devices that she classifies as being in
working order. (4 marks)

c) Let W ∈ [0, ∞) represent wind speed in miles per hour (mph). Assume
that W has the following density function:
2
f (w) = 2 λ w e−λ w
where λ > 0.
2
i) Show that Pr(W > w) = e−λw . (3 marks)

ii) If there is a 1% chance of wind speeds greater than 120mph, what is

the value of λ? (2 marks)

iii) A wind turbine only operates when the wind speed is between 10 mph
and 60 mph. Calculate the probability that the wind turbine does not
operate. Note: use λ value from part (ii). (3 marks)
Part I: Probability Question 2

a) Let X ∼ Geometric(p).

i) Derive the moment generating function for X:

et p
M(t) = .
1 − et (1 − p) (4 marks)

ii) Use the result in part (i) to derive E(X). (4 marks)

b) Clearly state the appropriate distribution, and then answer the associated
short question, in each of the parts below.

i) Flaws occur independently on a glass surface a rate of one per m2 .

Calculate the probability that six flaws occur on a 4m2 sheet of glass.
(3 marks)

ii) A manufacturing process yields components with weights contained

in the interval [1.15 kg, 1.28 kg] where all values are equally likely.
What proportion of components weigh less than 1.2 kg? (3 marks)

iii) Assume that 35% of people have the blood type A+. In a sample
of 100 randomly selected people, what is the probability that more
than 30 of them have this blood type? (3 marks)

c) Consider a game where a die is rolled and you win if a number divisible
by three appears. The game costs e10 to play and returns e25 (i.e., a
profit of e15) if you win.

You are thinking of two possible strategies:

• play the game 10 times, or
• keep playing until you win once.

Let X be the number of times you win, and let N be the number of times
you played the game. Calculate the expected profit under both of the
above strategies. (8 marks)
Part I: Probability Question 3

a) Let X ∼ Normal(µ = 20, σ = 3).

Calculate the following:

i) Pr(X < 25) ; (3 marks)

ii) Pr(23.5 < X < 28.4) ; (3 marks)

iii) the value x such that Pr(X > x) = 0.35 ; (3 marks)

iv) Pr(X1 + X2 > 45.7) where X1 and X2 are both independent

Normal(µ = 20, σ = 3) random variables. (3 marks)

b) For a sample of independent random variables, X1 , . . . , Xn , with E(Xi ) = µ

and V ar(Xi ) = σ 2 , it can be shown that
 
X1 + · · · + Xn σ
X= ∼ Normal µ, √
n n

when n is large; this is the Central Limit Theorem.

Briefly explain how the above result leads to the 95% confidence interval
σ
X ± 1.96 √
n

and why this interval is useful in practice. (4 marks)

c) Consider the following joint distribution:

x
p(x, y) 0 1
0 0.4 0.2
y
1 0.1 0.3

i) Calculate E(X), E(X 2 ), E(Y ) and E(Y 2 ). (4 marks)

ii) Calculate Corr(X, Y ) and interpret its value. (5 marks)

Part II: Statistics Question 1

a) An engine manufacturer runs a sample of engines through a series of

diagnostic tests. Below is the number of tests failed by each of the engines
tested.

1 2 2 4 7 7 9 10 10 14
15 16 22 23 30 30 33 34 42 54
56 57 61 129 143

i) Construct a frequency table with 5 classes (note: let “0” be the lower
limit of the first interval). (4 marks)

ii) Draw the histogram. (2 marks)

iii) Calculate the mean and median for this sample. State which measure
is more appropriate (provide justification). (3 marks)

b) Briefly answer the following question. In terms of population parameters

of interest, what is the purpose of calculating statistics such as a sample
mean or a sample proportion? (3 marks)

c) It is believed that 55% of individuals will vote for a particular candidate

in an upcoming election. A sample of 250 individuals were contacted and
it was found that 122 plan to vote for the candidate in question.

i) What type of data was collected on each individual? (2 marks)

ii) What is the parameter here? (provide symbol and value) (2 marks)

iii) What is the statistic here? (provide symbol and value) (2 marks)

iv) Calculate a 99% confidence interval for the parameter; does this
support the researcher’s belief? (4 marks)

v) What sample size is required to reduce the margin of error in the

previous confidence interval to ± 0.005? (3 marks)
Part II: Statistics Question 2

a) Consider the following sample of data:

5 2 2 3 1 3

For this sample, calculate the following:

i) the mean ; (1 mark)

ii) the standard deviation ; (3 marks)

iii) an 80% confidence interval, and interpret the result. (4 marks)

b) A manufacturer wants to compare two designs of CPU in terms of their

normal operating temperature (in degrees Celsius) where lower is better.
Two small samples are selected and the results are as follows:

Design A Design B

sample size 10 12
mean 22.2 ◦ C 21.4 ◦ C
standard deviation 3.1 ◦ C 1.4 ◦ C

i) Before comparing means, the F test was carried out, and yielded a
p-value of 0.0161.

State the null and alternative hypotheses for this test, and your
conclusion based on the above p-value reported. (3 marks)

(Note: your conclusion here impacts your calculations in part (ii))

ii) Test the hypothesis that there is no difference in the mean operating
temperatures for the two designs using the 5% level of significance,
and provide your conclusion in both statistical and non-statistical
language. (11 marks)

iii) Based on the results of parts (i) and (ii), would you recommend either
Design A or Design B – or are both equivalent? (3 marks)
Part II: Statistics Question 3

A series of laboratory experiments were carried out to investigate the relationship

between the voltage level applied to an electronic component, and the lifetime
(in weeks) of the component. In particular, voltages in the set {5, 6, 7, 8, 9, 10}
were considered and, for each voltage, six components were tested until
failure. Thus, the total sample size was n = 36 components.
We define the following:
• L = the lifetime of the component in weeks,
• Y = ln(L) the natural logarithm of the lifetime, and
• X = the voltage level.
Furthermore, the following quantities have been calculated for you:
P P 2 P
Yi = 4.6 Yi = 177.1 Xi Yi = −99.4
P P 2
Xi = 270 Xi = 2130

Lifetime versus Voltage ln(Lifetime) versus Voltage

4
60
50

Y = ln(Lifetime)
2
L = Lifetime
40
30

0
20

−2
10
0

5 6 7 8 9 10 5 6 7 8 9 10
X = Voltage X = Voltage

a) Show that the correlation between X and Y is -0.98, and comment on

this value. Would you expect the magnitude of the correlation between
X and L to be greater than or less than this? Make reference to the above
scatterplots in your answer. (4 marks)

b) Estimate the coefficients of the regression equation Y = β0 + β1 X, and

interpret the slope of this line. (5 marks)

c) Comment on how well the regression model fits the data based on its
coefficient of determination. (2 marks)

d) Show that, from the equation in part (b), one can obtain the equation
L = L0 ω X . Hence, or otherwise, predict the lifetime of the component
when the voltage is 2V . Explain whether or not you believe the prediction
to be accurate, and why. (4 marks)

e) By computing a 95% confidence interval for β1 , test the hypothesis that

there is no relationship between log-lifetime and voltage level. Clearly
state the hypotheses and provide your conclusion in both statistical and
non-statistical language. (7 marks)

f) Using the confidence interval from part (e), produce a confidence interval
for ω (where ω appears in the model in part (d)). (3 marks)
Useful Formulae: Page 1
Histogram:
max(x) − min(x)
• class width =
number of classes

Numerical Summaries: P
xi
• x̄ =
n

x2i − n x̄2
P
2
• s =
n−1

n+1
• Position of Qk : ×k
4

• IQR = Q3 − Q1

• LF = Q1 − 1.5 × IQR

• UF = Q3 + 1.5 × IQR

Probability:
• Pr(Ac ) = 1 − Pr(A)

• Pr(A ∪ B) = Pr(A) + Pr(B) − Pr(A ∩ B)

• Pr(E1 ∪ E2 ∪ · · · ∪ Ek ) = Pr(E1 ) + Pr(E2 ) + · · · + Pr(Ek ) (if mutually exclusive)

• Pr(A ∩ B) = Pr(A) Pr(B | A) = Pr(B) Pr(A | B)

• Pr(E1 ∩ E2 ∩ · · · ∩ Ek ) = Pr(E1 ) Pr(E2 ) · · · Pr(Ek ) (if independent)

Pr(A ∩ B) Pr(A) Pr(B | A)

• Pr(A | B) = =
Pr(B) Pr(B)

• Pr(B) = Pr(B ∩ E1 ) + Pr(B ∩ E2 ) + · · · + Pr(B ∩ Ek )

= Pr(E1 ) Pr(B | E1 ) + Pr(E2 ) Pr(B | E2 ) + · · · + Pr(Ek ) Pr(B | Ek )
(if E1 , . . . , Ek are mutually exclusive & exhaustive)
 Useful Formulae: Page 2
Counting Techniques:
• n ! = n × (n − 1) × (n − 2) × · · · × 3 × 2 × 1
 
n n!
• =
k k ! (n − k) !

Random Variables:
P R
• E(X) = xi p(xi ) or x f (x) dx

• E(aX + b) = a E(X) + b

• E(X + Y ) = E(X) + E(Y )

• E(X 2 ) = x2i p(xi ) or x2 f (x) dx

P R

P P RR
• E(XY ) = i j xi yj p(xi , yj ) or x y f (x, y) dx dy

• E(XY ) = E(X)E(Y ) if X and Y are independent

• V ar(X) = E[ ( X − E(X) )2 ] = E(X 2 ) − [E(X)]2

• V ar(aX + b) = a2 V ar(X)
p
• Sd(X) = V ar(X)

• Cov(X, Y ) = E[ ( X − E(X) ) ( Y − E(Y ) ) ] = E(XY ) − E(X)E(Y )

Cov(X, Y )
• Corr(X, Y ) = p
V ar(X) V ar(Y )

Discrete Distributions:

• X ∼ Binomial(n, p) • X ∼ Geometric(p) • X ∼ Poisson(λ)

• x ∈ {0, 1, 2, . . . , n} • x ∈ {1, 2, . . . , ∞} • x ∈ {0, 1, 2, . . . , ∞}

n λx
px (1 − p)n−x • p(x) = (1 − p)x−1 p


• p(x) = x
• p(x) = x!
e−λ

• Pr(X ≥ x) = tables • Pr(X ≥ x) = (1 − p)x−1 • Pr(X ≥ x) = tables

1
• E(X) = n p • E(X) = p
• E(X) = λ
1−p
• V ar(X) = n p (1 − p) • V ar(T ) = p2
• V ar(X) = λ
 Useful Formulae: Page 3
Continuous Distributions:

• T ∼ Exponential(λ) • X ∼ Uniform(a, b) • X ∼ Normal(µ, σ)

• t ∈ [0, ∞) • x ∈ [a, b] • x ∈ (−∞, ∞)

1
• f (t) = λ e−λ t • f (x) = b−a
• f (x) = not needed

b−x
• Pr(T > t) = e−λ t • Pr(X > x) = b−a
• Pr(X > x) = tables

1 b+a
• E(T ) = λ
• E(X) = 2
• E(X) = µ
1 (b−a)2
• V ar(T ) = λ2
• V ar(X) = 12
• V ar(X) = σ 2

Properties of Normal Distribution:

• (1 − α)100% of the Normal(µ, σ) distribution lies in µ ± z α/2 σ
 
x−µ
• Pr(X > x) = Pr Z >
σ

• Z ∼ Normal(µ = 0, σ = 1)

• Pr(Z < −z) = Pr(Z > z)

• Pr(Z > −z) = Pr(Z < z) = 1 − Pr(Z > z)

• If X1 ∼ Normal(µ1 , σ1 ) and X2 ∼ Normal(µ2 , σ2 )

 q 
2 2
⇒ Sum: X1 + X2 ∼ Normal µ1 + µ2 , σ1 + σ2

 q 
2 2
⇒ Difference: X1 − X2 ∼ Normal µ1 − µ2 , σ1 + σ2

p
• For X1 , . . . , Xn ∼ any distribution with µ = E(X) and σ = Sd(X) = V ar(X)
 
σ
⇒ Sample mean: X ∼ Normal µ, √ if n > 30
n
 Useful Formulae: Page 4
Statistics and Standard Errors:

Parameter Statistic Standard Error Samples Details

s
µ x̄ √ large / small ν =n−1
n
r
p p̂ p̂ (1 − p̂) large confidence
interval
n
r
p0 (1 − p0 ) large hypothesis
test
n

(a + b)2
µ1 − µ2 x̄1 − x̄2 s large / small ν = a2
s21 s2
2
+ n2b−1
+ 2 n1 −1
n1 n2
s21 s2
a= , b= 2
n1 n2
s
s2p s2p
+ small ν = n1 + n2 − 2
n1 n2
assuming
(n1 − 1) s21 + (n2 − 1) s22
where s2p = σ12 = σ22
n1 + n2 − 2
s
p1 − p2 p̂1 − p̂2 p̂1 (1 − p̂1 ) p̂2 (1 − p̂2 ) large confidence
+ interval
n1 n2
s
p̂c (1 − p̂c ) p̂c (1 − p̂c )
+ large hypothesis
n1 n2 test
x1 + x2
where p̂c =
n1 + n2

Confidence Intervals:

• Large sample: statistic ± z α/2 × standard error

• Small sample: statistic ± t ν, α/2 × standard error

Useful Formulae: Page 5
Hypothesis Testing:
statistic − hypothesised value
• z=
standard error



 2 × Pr(Z > |z|) if Ha : µ 6= µ0


• p-value = Pr(Z < z) if Ha : µ < µ0



Pr(Z > z) if Ha : µ > µ0


larger variance s2larger

• F = = 2
smaller variance ssmaller

ν1 = n top − 1, ν2 = n bottom − 1

Correlation and Regression:

P
( xi )2
• Sxx = x2i −
P
n

P
( yi )2
yi2 −
P
• Syy = n

P P
P ( xi )( yi )
• Sxy = xi yi − n

Sxy
• r=p
Sxx Syy

• β̂0 = ȳ − β̂1 x̄

Sxy
• β̂1 =
Sxx

e2i
P
Syy − β̂1 Sxy
• s2e = =
n−2 n−2
s
s2e
• s(β̂1 ) =
Sxx

• ν = n − 2 (degrees of freedom)

• R2 = r 2