ISYE 6501 ADVANCED ANALYTICS
MODELING COMPREHENSIVE FINAL
EXAM QUESTIONS AND ANSWERS
1. When using Support Vector Machines (SVM), what does the introduction of slack variables
(C parameter) specifically allow for?
A. The use of non-linear kernels like RBF
B. Automatic selection of the most important variables
C. Increasing the number of features to handle high-dimensional data
D. Permitting some data points to be on the wrong side of the margin or hyperplane
Answer: D
Conceptual Explanation: In SVM, slack variables are used to create a soft margin, which
allows for some misclassifications to achieve a more robust overall model.
2. In K-means clustering, which technique is most commonly used to determine the optimal
number of clusters by observing the decrease in within-cluster sum of squares?
A. The Silhouette Method
B. Dendrogram Pruning
C. Cross-validation
,D. The Elbow Method
Answer: D
Conceptual Explanation: The Elbow Method plots the WCSS against the number of
clusters; the ‘elbow’ point indicates where adding more clusters provides diminishing
returns in variance explanation.
3. Which of the following is a key difference between Lasso (L1) and Ridge (L2) regression?
A. Lasso regression can perform feature selection by shrinking coefficients to zero
B. Lasso regression is more effective when all features are relevant
C. Ridge regression can set coefficients exactly to zero
D. Ridge regression is only used for classification tasks
Answer: A
Conceptual Explanation: Lasso uses the L1 penalty, which has the mathematical property
of being able to force coefficients to exactly zero, effectively performing variable selection.
4. If you are modeling a time series with a trend and a seasonal pattern where the seasonal
variations grow over time, which Holt-Winters model is most appropriate?
A. Simple Exponential Smoothing
B. Additive Holt-Winters
C. Multiplicative Holt-Winters
D. Double Exponential Smoothing
, Answer: C
Conceptual Explanation: The multiplicative version is used when the seasonal component
changes in proportion to the level (mean) of the series.
5. In the context of optimization, what does a ‘dual price’ (or shadow price) represent for a
constraint?
A. The cost of doubling the constraint’s limit
B. The penalty for violating the constraint
C. The change in the objective function value per unit increase in the constraint’s right-
hand side
D. The equilibrium price where supply equals demand
Answer: C
Conceptual Explanation: The dual price indicates how much the objective function would
improve if the constraint were relaxed by one unit.
6. Which test is specifically designed to detect a single outlier in a dataset that is assumed to
follow a normal distribution?
A. CUSUM Test
B. Chi-Square Test
C. Grubbs’ Test
D. Box-Cox Test
MODELING COMPREHENSIVE FINAL
EXAM QUESTIONS AND ANSWERS
1. When using Support Vector Machines (SVM), what does the introduction of slack variables
(C parameter) specifically allow for?
A. The use of non-linear kernels like RBF
B. Automatic selection of the most important variables
C. Increasing the number of features to handle high-dimensional data
D. Permitting some data points to be on the wrong side of the margin or hyperplane
Answer: D
Conceptual Explanation: In SVM, slack variables are used to create a soft margin, which
allows for some misclassifications to achieve a more robust overall model.
2. In K-means clustering, which technique is most commonly used to determine the optimal
number of clusters by observing the decrease in within-cluster sum of squares?
A. The Silhouette Method
B. Dendrogram Pruning
C. Cross-validation
,D. The Elbow Method
Answer: D
Conceptual Explanation: The Elbow Method plots the WCSS against the number of
clusters; the ‘elbow’ point indicates where adding more clusters provides diminishing
returns in variance explanation.
3. Which of the following is a key difference between Lasso (L1) and Ridge (L2) regression?
A. Lasso regression can perform feature selection by shrinking coefficients to zero
B. Lasso regression is more effective when all features are relevant
C. Ridge regression can set coefficients exactly to zero
D. Ridge regression is only used for classification tasks
Answer: A
Conceptual Explanation: Lasso uses the L1 penalty, which has the mathematical property
of being able to force coefficients to exactly zero, effectively performing variable selection.
4. If you are modeling a time series with a trend and a seasonal pattern where the seasonal
variations grow over time, which Holt-Winters model is most appropriate?
A. Simple Exponential Smoothing
B. Additive Holt-Winters
C. Multiplicative Holt-Winters
D. Double Exponential Smoothing
, Answer: C
Conceptual Explanation: The multiplicative version is used when the seasonal component
changes in proportion to the level (mean) of the series.
5. In the context of optimization, what does a ‘dual price’ (or shadow price) represent for a
constraint?
A. The cost of doubling the constraint’s limit
B. The penalty for violating the constraint
C. The change in the objective function value per unit increase in the constraint’s right-
hand side
D. The equilibrium price where supply equals demand
Answer: C
Conceptual Explanation: The dual price indicates how much the objective function would
improve if the constraint were relaxed by one unit.
6. Which test is specifically designed to detect a single outlier in a dataset that is assumed to
follow a normal distribution?
A. CUSUM Test
B. Chi-Square Test
C. Grubbs’ Test
D. Box-Cox Test
