You are on page 1of 4

UNIVERSITY OF LIMPOPO

FACULTY OF SCIENCE AND AGRICULTURE

SCHOOL OF MATHEMATICAL AND COMPUTER SCIENCES

TUTORIAL 4

SECTION A

a) In the matrix form of the simple linear regression mode, the least squares estimator for  is
β  X X X Y where the element of X are fixed constants in laboratory experiment.


Derive E β̂ and Var β̂ . 
b) For the multiple regression model, with an intercept term, complete the following part,
ˆ e
YY ˆ  Xβˆ
Y

ˆ ˆ
prove that Y  YY  ee .
SECTION B

In an effort to decide whether or not to apply for a liquor license, a restaurant owner sampled ten
comparable restaurants in his town and recorded the total profits (Y), the food sales (X 1), and liquor
sales (X2) for the past month. His sample yielded the following results. (All figures are in thousands
rands).
y i =β 0 + β 1 x i1 + β 2 x i2 + ε i
The linear model where  is the error term was proposed.

; ;

[ ]
31 . 4
X ' Y = 86 . 58
101 .37 and ∑ y 2= 107 .14
TUTORIAL 4 2024

β1 β2 .
a) Find the multiple regression equation and interpret the estimates of the and
b) Predict the total profits in a restaurant with R1700 food sales and with R3000 liquor

sales.
c) Test for the overall.
d) Determine the coefficient of multiple determination and interpret its value.

e) At the 5% level of significance, determine whether each independent variable makes a


significant contribution to the regression model. On the basis of these results, indicate the
regression model that should be used in this problem.

f) Repeat e) using confidence level approach.

g) Construct and interpret a 95% confidence interval of the average profits in a restaurant with
R1700 food sales and with R3000 liquor sales.

h) Construct and interpret partial coefficient of determination.

SECTION C

The data consist of the 67 houses with Y = Sales Price of the house (“salesprice” on the output in
Table 2), and the three predictor variables are:

Y=Sales Price of the house in rands

X1 = Square Feet divided by 100 = “sqft100”


X2 = Number of bedrooms = “bedrooms”
X3 = Lot Size in square feet= “lotsize”

The model used did not involve any transformations; it is


E{Yi} = β0 + β1Xi1 + β2Xi2 + β3Xi3.
The results/output analyzed by a statistical package is given by

2|Page
TUTORIAL 4 2024

Table 1

Source | SS df MS Number of obs = 67


-------------+------------------------------ F = 6.69
Model | 2.4910e+11 3 8.3035e+10 Prob > F = 0.0005
Residual | 7.8158e+11 63 1.2406e+10 R-squared = 0.2417
-------------+------------------------------ Adj R-squared = 0.2056
Total | 1.0307e+12 66 1.5616e+10 Root MSE = 1.1e+05

------------------------------------------------------------------------------
salesprice | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
sqft100 | 9893.524 2508.561 3.94 0.000 4880.564 14906.48
bedrooms | -36018.68 14665.17 -2.46 0.017 -65324.68 -6712.675
lotsize | 2.512511 1.124531 2.23 0.029 .2653148 4.759707
intercept | 290558.1 75398.79 3.85 0.000 139885.6 441230.5
------------------------------------------------------------------------------

Sequential Sum of Squares for Regression

NOTE: In case you aren’t familiar with this notation, 1.18e+11 = 1.18 x 1011, as an example.

a) The estimated regression equation for the full model with all 3 variables, filling in numbers for
the coefficients is given by
Y^ = 290558.1 + 9893.524X1 – 36018.68X2 + 2.513X3.
Interpret the values of the three regression coefficients.

b) Is the overall model useful? Carry out an appropriate test and draw a conclusion using
α =0 . 05 .
c) Test for the contribution of each independent variable using p-value approach.
d) The 95% confidence interval for the intercept, β0, is given by (139885.6; 441230.5).
Draw conclusions about the tests of the other three coefficients using their confidence
intervals.
SECTION D
A statistician tried to fit a model to house prices as a function of the following variables as given in
Table 2:
Table 2
Tax Property taxes in thousands of rands
Age Age of the house in years
Bed Number of Bedrooms
Bath Number of Bathrooms
Space Living space in thousands of square metre

3|Page
TUTORIAL 4 2024

Lot Lot size in thousands of square metre

The response variable was the sale price of the house in thousands of rands. Below is an
ANOVA table from fitting this model.

Analysis of Variance
Sum of Mean
Source DF Squares Square F Value
Model a 675.85947 d f
Error b c e
Corrected Total 23 831.50958

(a) Give the complete ANOVA table by filling in the missing values a, b, c, d, e and f.
(b) State the null and alternative hypothesis that the F value (f) is testing. Give the conclusion of the
test for this fitted model at the significance level of 0.05.

4|Page

You might also like