Download as pdf or txt
Download as pdf or txt
You are on page 1of 2

CHRIST (DEEMED TO BE UNIVERSITY),Bengaluru - 560029

End Semester Examination May/June - 2022


MSc MATHEMATICS II SEMESTER
Code: MTH443A Max.Marks: 100
Course: REGRESSION ANALYSIS Duration: 3Hrs
SECTION A
Answer all relevent questions 5X20=100
1 a) List the different uses of regression analysis with suitable (10)
examples.
[OR]
2 b) i) Explain the assumptions of a simple linear regression model. (10)
ii) Derive least square estimators of parameters involved in a
simple linear regression model.
3 a) Explain the properties of least square estimators of β 0and β 1 for (10)
a simple linear regression model y i = β 0 + β 1 x i + ϵ i , i = 1, 2, . . . , n.
[OR]
4 b) In fitting a simple linear regression model y i = β 0 + β 1x i + ϵ i, i = 1,(10)
2, . . . , n; E(
, construct 99% confidence intervals for β 0,β 1 and σ .2

5 a) It is thought that the number of cans damaged in a boxcar (10)


shipment of cans is a linear function of the speed of the boxcar
at impact. Thirteen boxcars selected at random were used to
examine whether this appeared to be true. The data collected were
as follows:
Speed of Car at impact Number of cans Damaged
3 54
5 86
8 136
4 65
3 109
3 28
4 75
3 53
5 33
7 168
3 47
8 52
i) Determine a straight-line relationship is sensible? (Use α = 0.05.
)
ii) Obtain a 99% confidence interval for the slope of the
regression line.
[OR]
6 b) Explain the multiple linear regression model using matrix (10)
theoretic approach and state its assumptions.
7 a) Find the least-squares estimators of parameters involved in the (10)
multiple linear regression model with k regressors by matrix
theoretic approach. Are they unbiased?
[OR]
8 b) Let Y i = β 0 + β 1X 1i + β 2X 2i + ϵ i, (i = 1, 2, 3), where Y i = {5, 3, 4}, X 1i = {1,
(10)4, 2}
and X 2i = {2, 1, 0}respectively. Find the least square estimates for
β 0 , β 1and β 2. Also obtain the residuals associated with them.
9 a) The fitted muliple linear model for a data of 25 observations on (10)
response variable y and 2 regressors is given by ŷ = 2.341 + 1.615β 1 + 0.014β 2
. The sum of squares due to residuals is found to be 233.726 and
diagonal elements of (X TX) − 1 is given by 0.11321, 0.00274378 and 0.00000123
. Construct 99% confidence intervals for all the parameters
involved in the model.

[OR]
10 b) i) Define R2 and adjusted R2in the case multiple linear (10)
regression model with k regressors. Suppose the computed value of
adjusted R2 is 0.15. How will you interpret the model?
iii) Suppose a new regressor is added to the model, then what
will happen to the value of R2 and adjusted R square.
11 a) What is the need for the constant variance of the error term in (10)
regression model building? Discuss various variable
transformations to remove non- constant variance.
[OR]
12 b) Discuss various variable selection procedures for building a (10)
linear regression model and which among these is best procedure.
13 a) Explain i) Studentized residuals ii) Standardized residuals iii) (10)
PRESS residuals iv) How to detect outliers using residuals.
[OR]
14 b) Discuss the use of residual plots in regression analysis. (10)
15 a) Suppose a researcher wants to analyze a data with one response (10)
variable and k number of regressors. What all methods you will
suggest to diagnoses the multicollinearity in the data? and what
are all the remedies you can suggest to remove multicollinearity.
[OR]
16 b) (i) Explain The term autocorrelation. (10)
(ii) Discuss the Durbin-Watson test to detect autocorrelation.
17 a) Explain the method of least squares in non-linear regression. (10)
[OR]
18 b) What do you mean by logistic regression. Suggest a model suitable (10)
for binary response data. Explain?
19 a) How will you model the categorical response variable? Discuss. (10)
[OR]
20 b) Suppose the response variable is count data, what regression (10)
model you will suggest for regression analysis. Explain?

You might also like