0% found this document useful (0 votes)
6 views2 pages

Tutorial01 Solution

Uploaded by

Shen Tiffany
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
6 views2 pages

Tutorial01 Solution

Uploaded by

Shen Tiffany
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

THE HONG KONG POLYTECHNIC UNIVERSITY

DEPARTMENT OF APPLIED MATHEMATICS


AMA3602 APPLIED LINEAR MODELS

Tutorial 1 Solution

1. (a) Notice that X1 ∼ N(0, 1) and 12 2 2


P
i=2 Xi ∼ χ11 . P
Since X1 is independent of
X2 , . . . , X12 , it follows that X1 is independent of 12 2
i=2 Xi . Thus,

X1 11X1
qP = qP = Y1 ∼ t11 ,
12 2 12 2
i=2 X i /11 i=2 X i

where t11 denote a random variable following t-distribution with 11 degrees of


freedom. From the t-table, we find that P (Y1 = t11 > 1.363) = 0.1.
(b) Notice that 3i=1 Xi2 ∼ χ23 and 12 2 2
P P
i ∼ χ9 . Since X1 , X2 , X3 P
i=4 XP are inde-
pendent of X4 , . . . , X12 , it follows that i=1 Xi2 is independent of 12
3 2
i=4 Xi .
Thus,
P3 2
3 3i=1 Xi2
P
i=1 Xi /3
P12 2
= P12 2
= 3Y2 ∼ F3,9 ,
i=4 X i /9 i=4 X i

where t3,9 denote a random variable following F -distribution with degrees of


freedom 3 and 9. In addition, we have

P (Y2 > c) = P (3Y2 = F3,9 > 3c) = 0.01.

From the F -table, we find that 3c = 6.99 and hence, c = 2.33.

2. (a) Since the alternative hypothesis is µ ̸= 0, we conduct a two-sided test. Under


the null hypothesis µ = 0, the test statistic T ∼ N (0, 10). Therefore, the
rejection region is
 
T −0
T : √ > z0.05/2 = 1.96 ,
10
where zα is a constant satisfying P (Z > zα ) = α for Z ∼ N(0, 1).

(b) Under the null hypothesis, the statistic (T − 0)/ 10 ∼ N(0, 1). We obtain

p-value = P (|Z| > |tobs / 10| ≈ 1.581) = 2P (Z > 1.581) ≈ 2 × 0.0571 ≈ 0.1142.

Since p-value > 0.05, we do not reject H0 at the 5% significance level.

3. (a) False. The intercept and slope of the true model are constants.
(b) True. The expected value of Y given X is

E(Y | X) = E(1 + 2X + ε | X) = 1 + 2X + E(ε | X) = 1 + 2X + 0.

1
(c) False. Since ε is random, P (Y = 1 + 2(2) + ε = 5 | X = 2) = P (ε =
0 | X = 2) < 1. In particular if ε follows normal distribution, we have
P (Y = 5 | X = 2) = 0.
(d) True. Conditioning on X, the covariance of Y and ε is

Cov(Y, ε | X) = Cov(1 + 2X + ε, ε | X) = Cov(ε, ε | X) = Var(ε) = 3.

4. (a) The intercept is the expected hourly wage when the years of working expe-
rience is 0, i.e. E(Y |X = 0). The slope represents the change in expected
hourly wage associated with a one-unit increase in years of working experience.
(b) The expected hourly wage for a person with 10 years of working experience is

E(Y | X = 10) = 5.3733 + 0.0307(10) = 5.6803.

(c) Let Y and X denote the hourly wage and years of working experience of this
person respectively. We have

P (Y ≥ 6 | X) = 0.5
P (5.3733 + 0.0307X + ε ≥ 6 | X) = 0.5
P (ε ≥ 0.6267 − 0.0307X | X) = 0.5.

Since ε ∼ N(0, 13.49) is independent of X, this implies


 
ε
P √ ≥ 0.6267 − 0.0307X X = P (Z ≥ 0.6267 − 0.0307X) = 0.5,
13.49
where Z ∼ N(0, 1). Remark that P (Z ≥ 0) = 0.5. Thus,

0.6267 − 0.0307X = 0
X ≈ 20.414.

The minimum number of years of working experience required is 20.42.


(d) Let Ye denote the hourly wage in Hong Kong dollar. Since 1 US dollar = 7.78
Hong Kong dollar, we obtain

Ye = 7.78Y
= 7.78(5.3733 + 0.0307X + ε)
= 41.80427 + 0.238846X + 7.78ε.

Let εe = 7.78ε. Thus, the linear model in Hong Kong dollars is:

Ye = 41.80427 + 0.238846X + εe,

where E(e ε) = 7.782 × 13.49 = 816.5281, and εe is independent of


ε) = 0, Var(e
X.

You might also like