FUND. OF MATH.

& DATA ANALYSIS TOOLS (17/01/23)

Surnames, First name ...........................................................................................................................................

OBSERVATION: There is only one correct answer per question. Each correct answer will be scored with 1 point and each incorrect question will score -1/3.
Unanswered questions will not be scored. The chosen answers that will be considered valid are those listed in the table attached below.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
c b a c c a a b c b b c a c b b

1. The following figure shows the number of customers vi- c) The percentage of variance of the stock explained by
siting a small fashion shop over several weeks. From the the model is around the 30 %.
inspection of the figure it can be deduced that: Solution: In view of the results, the fitted linear regres-
sion line would be:
Stock = 611.368404 − 0.047564 · Sales.
Since the slope is negative (−0.047564), as the number
of sales increases the stock of the shop decreases, so the
relationship between the variables is inverse.
The coefficient of determination, R2 , is equal to 0.7063.
Moreover, we know that it coincides with the square of
the linear correlation coefficient (r):
R2 = 0.7063 = r2 .
Therefore, taking into accounte that the slope is negative,
the linear correlation coefficient is:
√
a) The day with most customers was a Saturday. r = − 0.7063 = −0.8404.
b) The distribution of the number of customers visiting
the fashion shop on Friday is negative asymmetric.
c) The median of the number of customers on Saturday 3. Next figure shows the histograms of daily energy consum-
is lower than the third quartile of the number of cus- ption in a clothing store during 2019. Different histograms
tomers on Thursday. have been obtained for Weekdays, Saturdays and Sundays.
Which of the following options is true?.
2. A linear regression model has been computed relating the
stock and the number of sales of a shop with data for the
year 2022. The result obtained with R is the following:
Call:
lm(formula = Stock ~ Sales, data = shop)

Residuals:
Min 1Q Median 3Q Max
-2159.5 -414.9 -106.1 277.5 3789.5

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 611.368404 134.926030 4.531 0.0000228 ***
Sales -0.047564 0.003614 13.160 < 2e-16 ***
---
Signif. codes: 0 ’***’ 0.001 ’**’ 0.01 ’*’ 0.05
’.’ 0.1 ’ ’ 1
Residual standard error: 814.5 on 72 degrees of freedom
(1 observation deleted due to missingness)
Multiple R-squared: 0.7063,Adjusted R-squared: 0.7023 a) The distribution of energy consumption on Weekdays
F-statistic: 173.2 on 1 and 72 DF, p-value: < 2.2e-16 is asymmetric to the right.
Taking this output into account, it can be stated that: b) The high values of energy consumption are more fre-
a) The linear relationship between the variables is strong quent than those moderate and low, for all the days.
and direct. c) The distribution of energy consumption on Sunday is
b) The linear correlation coefficient between the two va- not asymmetric to the left.
riables is equal to −0.8404.
4. The following observations correspond to the total num- Furthermore:
ber of subscribers to a fashion magazine from year 2016
P (I ∪ T ) = P (I) + P (T ) − P (I ∩ T ) = 0.4 + 0.33 − 0.2 = 0.53.
to 2021: 5175, 5205, 5200, 5165, 5125, 5122. The median
number of subscribers to the magazine is given by: Therefore:
a) 5182.5. P (I ∪ T ) = 1 − 0.53 = 0.47
b) 5165.33.
c) None of the other answers.
Solution: First, we have to sort the data:
7. In an alarm system of a building, the probability that the
5122, 5125, 5165, 5175, 5200, 5205 alarm signal is activated in the presence of danger is 0.95
As n = 6 is an odd number, the median is calculated as and the probability that it is activated in the absence of
the mean of the two central values: any danger is 0.03. Further, the probability that a dange-
5165 + 5175 rous situation happens is 0.1. What is the probability that
Me = = 5170 the alarm signal is activated?
2
a) 0.122.
b) 0.980.
5. Grades in a Mathematics course and a Marketing course c) 0.858.
taken simultaneously were as follows for a group of stu- Solution: Let us define the following events:
dents: A =“The alarm signal is activated”,
Marketing grades D =“Presence of a dangerous situation”,
Distinction Good Pass Fail We have the probabilities: P (A|D) = 0.95, P (A|D) =
Math grades 0.03, P (D) = 0.1. Note that, P (D) = 1−P (D) = 1−0.1 =
Distinction 25 6 17 13 0.9.
Good 17 16 15 6 Applying the theorem of Total Probabilities, we have:
Pass 18 4 18 10 P (A) = P (A|D)P (D) + P (A|D)P (D)
Fail 10 8 11 20
= 0.95 · 0.1 + 0.03 · 0.9
What is the relative frequency of students who did not fail = 0.122
the Marketing exam knowing that they failed the Mathe-
matics exam?
a) 13.55 %.
b) 63.55 %. 8. Concerning question 7, if the alarm signal has been ac-
c) 59.18 %. tivated, what is the probability that there happened a
Solution: dangerous situation?
a) 0.950.
Marketing grades
b) 0.779.
Distinction Good Pass Fail Total c) 0.882.
Math grades Solution: By Bayes’ theorem, we have:
Distinction 25 6 17 13 61
Good 17 16 15 6 54 P (A|D)P (D)
P (D|A) =
Pass 18 4 18 10 50 P (A|D)P (D) + P (A|D)P (D)
Fail 10 8 11 20 49 0.95 · 0.1
= = 0.7786885.
Total 70 34 61 49 214 0.95 · 0.1 + 0.03 · 0.9
P (not failing Marketing exam|Fail Math exam) =
10 + 8 + 11 9. A box of candies has many different colours in it. There is
= = 0.5918 = 59.18 %.
49 a 15 % chance of getting a pink candy. If ten candies are
randomly taken from the box with replacement, what is
6. A survey about TV on demand services asked to Genera- the probability that exactly 4 candies are pink?
tion Z reveals that 40 % have a Netflix subscription, 33 % a) 0.60.
have a HBO subscription and 20 % have a subscription b) 0.40.
to both services. If a zillenial is chosen at random, what c) 0.04.
is the probability that he/she does not have any of these Solution: Let us define the random variable X =“Number
services contracted? of pink candies in the box of 10”. Note that, X ∼ B(n =
a) 0.47. 10, p = 0.15).
b) 0.13. The probability that exactly four candies in the box of ten
c) 0.53. are pink:
Solution: Let us define the following events:
!
10
I =“Have a Netflix subscription” P (X = 4) = 0.154 (1 − 0.15)10−4
4
T =“Have a HBO subscription”.
Take into account the statement of the exercise, we know 10!
= 0.154 0.856
that: P (I) = 0.4, P (T ) = 0.33 and P (I ∩ T ) = 0.2. (10 − 4)!4!
If a zillenial is chosen randomly, the probability that he/she = 0.04.
does not have any service contracted is:
P (I ∪ T ) = 1 − P (I ∪ T )
10. A shop wants to carry out a draw of 10 gifts among the 20 Solution: Applying the Central Limit Theorem (or di-
customers who are shopping, there being 5 women and 15 rectly the properties of the Normal distribution) we have:
men. If the gifts are randomly given without replace- √
µ = 200 · 1.5 = 300, σ = 200 · 10.24 = 45.25
ment among the customers, what is the probability that
there are exactly 2 women among the winners?
a) 0.282.
b) 0.348. 14. The manager of a shop wants to analyse the waiting times
c) 0.5. of the customers to be attended. Information about 25
Solution: Let us define the random variable X =“Number customers is available, obtaining a sample mean of the
of women in the sample of 10 winners, extracted from a waiting time of 3 minutes with a standard quasi-deviation
population of 20’. Note that, X ∼ H(20, 10, 5). of 2 minutes. The 90 % confidence interval for the mean
The probability that there are exactly two women among of the waiting time of the customers is:
the winners is: a) (1.858, 4.142).
5 15


b) (2.473, 3.527).
P (X = 2) = 2 5
8 = 0.348. c) (2.316, 3.684).
2 Solution: Let us define: X =“Waiting time of a custo-
mer to be attended” that follows a normal distribution
N (µ, σ). We know that: n = 25, X̄ = 5 and Ŝ = 2.
11. A bakery sells chocolates birthday cakes through the in- We have to compute a confidence interval for µ (with unk-
ternet. Orders for chocolate cakes are random and arrive nown σ) for a confidence level of (1 − α)100 % = 90 %:
at the constant rate of 4.5 per day. At the start of any gi-    
ŝ 2
ven day, the bakery produces 6 cakes and produce no more X̄ ± t24,0.1/2 √ = 3 ± 1.711 √ = (2.3156, 3.6844),
until the day after. The probability that, by the middle of 25 25
a working day, the bakery would have sold half the cakes where t24,0.05 = 1.711 is taken from t-Student tables.
it produced for that day is:
a) 0.169. 15. Concerning question 14, what is the minimum sample size
b) 0.200. (number of customers) needed in order to estimate the
c) 0.667. mean of the waiting time with a 90 % confidence level
Solution: Let us define: X day
=“Number of cakes orders with an error length equal or less than 1?
in a day”. Note that, X day ∼ P oisson(λday = 4.5). The a) 12.
average number of cake orders in half of a day is: b) 11.
c) 44.
4.5
λday/2 = = 2.25 Solution:
2 !2
zα/2 · Ŝ
Then, X day/2
∼ P oisson(λ day/2
= 2.25). The probability n ≥
Error
we need to compute is:  2
1.645 · 2
e−2.25 2.253 ≥ = 10.8241
P (X = 3) = = 0.20 1
3!
Then, the minimum sample size needed must be n = 11.

16. The manager of a big fashion mall states that the pro-
12. The percentage of cotton in a garment follows a normal portion of customers who prefers to pay with credit card
distribution with mean 1.5 and variance 10.24. The pro- is greater than 40 %. To analyse this, a survey over 150
bability that a garment contains less than 2.8 percentage customers is done, observing that 50 prefer credit card
of cotton is: payment. For a significance level of α = 0.05, does the
a) 0.3409. corresponding hypothesis test support the manager state-
b) 0.9974. ment?
c) 0.6591. a) No, because p-value is equal to 0.9525.
Solution: The variable X =“Percentage of cotton √ in a b) Yes, because p-value is equal to 0.0475.
garment” follows a normal distribution N (1.5, 10.24 = c) Yes, because p-value is equal to -1.67.
3.2). Solution: This is a hypothesis test for a proportion p.
2.8 − 1.5 The hypothesis testing we should perform is:
P (X < 2.8) = P (Z < ) ≈ P (Z < 0.41)
3.2 H0 : p ≤ 0.4
= 1 − P (Z ≥ 0.41) = 1 − 0.3409 = 0.6591
H1 : p > 0.4
where Z ∼ N (0, 1). If H0 were true, we would have that
p̂ − 0.4
13. Regarding question 12, suppose that the garments are pac- D= p ∼ N (0, 1)
kaged in boxes of 200 units. If we take a box, what would 0.4(1 − 0.4)/n
be the distribution of the total percentage of cotton of the 50 1
We have a SRS of n = 100 and p̂ = = , so:
garments inside the box? 150 3
a) N (µ = 300, σ = 45.25). 1
− 0.4
b) N (µ = 200, σ = 10.24). 3
D̂ = p = −1.666
c) N (µ = 300, σ = 640). 0.4(1 − 0.4)/150
Now we compute the associated p−value:
p-value = ·P (D > D̂) = P (D > −1.67)
= P (D < 1.67) = 1 − P (D > 1.67)
= 1 − 0.0475 = 0.9525