Questions For CET
Questions For CET
Questions For CET
1. Define random variable. Give two examples. Give a mathematical expression for random
variable.
2. What is multivariate situation? Give two examples. Define a mathematical expression for
a variate.
3. What do you mean by modeling? Explain with examples the types of modeling
approaches.
4. Define data. Classify data as per scale of measurement. Give examples for each case.
8. A quality control engineer wants to select the best vendor amongst the locally available
vendors. A preliminary analysis of defects data screen out most of the vendors and two
vendors remain competitive. But the quality control engineer wants to be sure about the best
vendor or otherwise will go for both the vendors.
a) Convert the above-mentioned practical problem into a statistical problem (give details of
the procedure you would adopt).
b) What method do you adopt to select the best vendor and how do you do that?
9. A population is characterized by a variable X defined as N(10, 25). A sample of size 25 is
drawn at random from the population. What is the pdf of X ?
10. The covariance matrix of X = N 2 ( µ , Σ) is given below. Obtain its correlation matrix.
9 6
Σ =
6 25
11. Define statistical distance. A bivariate normal variable = X N 2 ( µ , Σ) has mean vector μT
= (3, 5) and covariance matrix Σ as given in question 10. Obtain the equation for
statistical distance.
12. Define Hotelling’s T2. State with two examples the uses of Hotelling’s T2.
13. X is N 2 ( µ , Σ) . Obtain its bivariate normal density function (pdf).
14. Show that for p = 2, Mahalanobis D2= (X – µ)′Σ-1 (X – µ) is an equation of an ellipse. If
the variables are independent and have equal variance, what will be the shape of
Mahalanobis D2?
15. A quality control engineer obtains the following observations of a process with two
variables (i = 4 and p=2):
2 12
8 9
X=
6 9
8 10
Specify the distribution of Hotelling T2.
2 12
8 9
X 4×2 =
6 9
8 10
Compute
The following questions (v – vii) assume that the population covariance matrix ∑ equal S.
X1 − X 2
vi) Obtain the distribution of
X2 − X3
vii) Which of the following variables are independent
a. X1 and X2
b. X2 and X3
c. (X1 , X2) and X3
6 9
X 3×2 = 10 6
8 3
(i) Obtain the multivariate statistics for X.
(ii) What is the sampling distribution of T2 in this case?
(iii) Obtain simultaneous confidence interval for µ(take α = 0.05)
20. A manufacturer produces light bulbs of a particular type. The duration of continuous
blowing in hours and intensity of light in lux are the two important quality characteristics
for use. The manufacturer claims the following:
(i) The average life of the bulbs produced is 2000 hours (continuous blowing),
and
(ii) The average intensity in lux is 100.
As a retailer of this bulb, how do you build a statistical model to test the manufacturer’s
claim?
21. Define the followings (to the point)
1. Normality
2. Homoscedasticity
3. Linearity
4. Uncorrelated error terms
23. The followings are sample data provided by a logistic company on the weight of
six shipments, the distances moved, and the damage that was incurred:
Observation no. 1 2 3 4 5 6
Weight (‘000 lbs) 5 4 2 1.5 3.4 5.2
Distance (‘000 km) 2 2.5 1.3 2.8 1.0 1.9
Damage (‘000 Rs) 20 15 10 11 16 23
24. The followings are sample statistics provided by a automobile company on the
production, the logistic cost, and the sales volume that was incurred:
5 40 10
( X ' X)-1 = 40 360 86
10 86 30
30
X ' Y = 260
53
and ANOVA TABLE (incomplete)
Model SS DF MS F P
Regression 20.989
Residual 9.011
Total 30.000
25. What is parameter testing? For the given data below, obtain a multiple regression
equation. Test the parameters of the model and comment on the result.
26. Define statistical distance. Name a few multivariate models that are based on statistical
distance.
27. Evaluate Hotelling’s T2, for testing Ho : µ ' = [ 7,11] , using the data
2 8 6 8
X' = .
12 9 9 10
Specify the distribution of T2 for this situation.Test Ho at α = 0.05 level. What
conclusion do you reach?
28. How do you assess casualty? Is path model a casual model? Enumerate the steps in path
modeling.Consider the following path diagram. Obtain normal equations.
τ1
Y1
β21
X1
τ1
Y2
29. Define causal and correlation relationships .For the following causal relationships, develop
path diagrams and structural equations.
i) X1, X2 Y1
ii) X1, X2 Y1
X2, Y1 Y2
iii) X1, X2 Y1
Y1, Y2 Y3
For question (ii), develop normal equations and complete correlation matrix in terms of
parameters.
30. Consider the following path model with hypothesized model parameters (standardized).
∈1
∈3
0.70 -0.80
Y1 Y3
X1
-0.80 -0.50
Y2
∈2
i) Write out the structural equation for each of the endogenous variables.
ii) Determine the complete correlation matrix.
iii) Decompose the correlation between Y3 and Y2 into
5. Correlation due to direct effects,
6. Correlation due to indirect effects,
7. Correlation due to common causes, and
8. Correlation due to correlated causes,
iv) Suppose the above model is correct, but instead the researcher believed in and estimated
the following model:
Y2 Y3 ∈
What conclusion would the researcher likely draw? Discuss the consequences of this
misspecification.
31. The following data set is obtained from a study conducted to determine the effect of job
stress on work safety perception. Two measurements (p=2), job stress and safety perception were
taken on two groups of people with sample size of 19. The mean vectors and covariance matrices
are as follows:
The management claims that there is no significant difference in job stress and safety
perception between the two groups of employees. Does the data substantiate the claim at
α = 0.01?
32. Define principal components (PCs). For the following population covariance matrix obtain
4 2
PCs. Σ =
2 9
33. Suppose the random variables X1, X2 and X3 have the covariance matrix
5 −4 0
Σ = −4 3 0
0 0 2
Obtain principal components. Prepare scree plot and determine the number of components to be
retained. What are the correlation coefficients between the PCs and the X variables? What
conclusions can you draw from this?
34. Determine the population principal components Y1 and Y2 for the covariance matrix given
below and also calculate the proportion of the total population variance explained by the first
principal component.
5 2
Σ =
3 2
35. How many principal components can be extracted when the population covariance matrix is
σ2 σ2p 0
2 1 1
Σ σ p σ 2
σ 2 p, − < p< ?
0 σ2p σ2 2 2
36. What is factor analysis? Differentiate between confirmatory factor model and exploratory
factor model.
Σ L* L*' + ψ, where=
37. What is factor rotation? Show that= L* LT ( L, L* , T , Σ, ψ represent their
usual meaning).
38. Define specific variance, commonality, factor loadings, and factor scores.
39. Show that after orthogonal rotation of factors, the estimated covariance matrix for the observed
variables remains unchanged.
40. Suppose the random variables X1, X2 and X3 have the covariance matrix
5 −4 0
Σ = −4 3 0
0 0 2
Assume m = 1 factor model, calculate the loading matrix L and matrix of specific variances
Ψ using principal component solution method.
Calculate communalities and interpret these quantities. What portion of the total variance is
explained by the common factor?
41. The following correlation matrix was obtained from a secondary source of data.
Attribute (Variable) X1 X2 X3 X4 X5
X1 1.00
X2 0.02 1.00
X3 0.96 0.13 1.00
X4 0.42 0.71 0.50 1.00
X5 0.01 0.85 0.11 0.79 1.00
The SPSS program run shows that the eigen-values for the 5 principal components are
2.85, 1.81, 0.20, 0.10 and 0.03.
- Construct a scree plot.
- How many factors should be retained?
- Compute correlation between factors and attributes.
42. For (41), the un-rotated two-factor solution is as below:
Obtain a rotated (orthogonal) factor solution (loadings) and name the factors.
43. The following data shows un-rotated factor loadings for two factors with five variables.
Obtain rotated factor loadings (must use graph paper) and comment on the result.
X1 0.50 0.80
X2 0.60 0.70
X3 0.90 -0.25
X4 0.80 -0.30
X5 0.60 -0.50
__________________________________________________________________
44. Write down the general equations for structural equations modeling. Show that the variance-
covariance matrix of the indicator variables in structural equations model is [the terms
represent their usual meanings]:
47. As a manufacturer of heavy earth moving machinery, your company requires some parts of
the machine to be purchased from different suppliers. Amongst the purchased items, ‘boom
assembly’ is the most important part. The key characteristics of the ‘boom assembly’ are
vertical (X1) and horizontal reach (X2), angle of rotation (X3), and load bearing capacity
(X4). Consider the case as a vender selection problem and obtain statistical models for
selecting the best vendor using:
a) Pair-wise comparison of vendors, and
b) Comparison amongst all the vendors simultaneously.
48. What do you mean by confidence region (CR)? The figure given below shows the
performance status (99.73% CR) of three different processes (l = 1, 2, 3) characterized by
X N 2 ( µl , Σ l ) .
Process A
Xl2
Process B
USLl2
Specification
region +
Process C
LSLl2
Xl1
LSLl1 USLl1
49. What ways MANOVA differs from ANOVA? What statistic is commonly used in testing
hypothesis in MANOVA? Define it.
50. The following observations on two responses are collected for three treatments.
x1
The observation vectors are
x2
4 3 7 4 5
Treatment 1: , , , ,
6 7 5 8 8
4 2 2
Treatment 2: , , ,
5 7 5
3 6 2 3
Treatment 3: , , , .
6 2 1 4