ISYE 6501 FINAL EXAM - ANALYTICS
MODELING QUESTIONS AND ANSWERS
1. In Support Vector Machines (SVM), what is the primary purpose of the ‘slack variables’ in a
soft-margin classification?
A. To ensure that the margin between the two classes is as small as possible
B. To transform the data into a higher-dimensional space using the kernel trick
C. To allow for some misclassifications or points within the margin to find a feasible
solution
D. To eliminate the need for a hyper-parameter like C
Answer: C
Conceptual Explanation: Slack variables allow the model to ignore certain constraints
(points) that would otherwise make the problem non-separable, balancing the margin size
and classification errors.
2. Why is scaling/normalization essential before applying the K-Nearest Neighbors (KNN)
algorithm?
A. To ensure that the model can handle categorical data without encoding
B. To increase the likelihood of the model being linearly separable
,C. To reduce the computational complexity of calculating Euclidean distance
D. To prevent variables with larger scales from dominating the distance calculation
Answer: D
Conceptual Explanation: KNN relies on distance metrics; if one feature has a much larger
range than others, it will disproportionately influence the distance and the resulting
classification.
3. Which of the following is a significant disadvantage of the K-means clustering algorithm?
A. It is computationally too slow for large datasets
B. It cannot be used for numerical data
C. It automatically determines the optimal number of clusters
D. The results are sensitive to the initial selection of centroids
Answer: D
Conceptual Explanation: K-means is a heuristic algorithm that may converge to a local
optimum depending on where the initial centroids are placed.
4. In a CUSUM (Cumulative Sum) control chart, what does the parameter ‘H’ represent?
A. The expected mean of the process
B. The slack value to ignore small shifts
C. The standard deviation of the observations
, D. The threshold limit that triggers an alarm when exceeded
Answer: D
Conceptual Explanation: In CUSUM, H is the threshold value. If the cumulative sum
exceeds H, the process is considered to have shifted (out of control).
5. What is the primary objective of Principal Component Analysis (PCA)?
A. To maximize the correlation between the target variable and features
B. To reduce dimensionality while preserving as much variance as possible
C. To ensure that all features have a mean of zero and variance of one
D. To find the best linear regression fit for a dataset
Answer: B
Conceptual Explanation: PCA transforms a large set of variables into a smaller one that
still contains most of the original information (variance).
6. Which component of a time series represents long-term fluctuations that are not related to
calendar-based cycles?
A. Seasonality
B. Trend
C. Cyclicity
D. Random Error
MODELING QUESTIONS AND ANSWERS
1. In Support Vector Machines (SVM), what is the primary purpose of the ‘slack variables’ in a
soft-margin classification?
A. To ensure that the margin between the two classes is as small as possible
B. To transform the data into a higher-dimensional space using the kernel trick
C. To allow for some misclassifications or points within the margin to find a feasible
solution
D. To eliminate the need for a hyper-parameter like C
Answer: C
Conceptual Explanation: Slack variables allow the model to ignore certain constraints
(points) that would otherwise make the problem non-separable, balancing the margin size
and classification errors.
2. Why is scaling/normalization essential before applying the K-Nearest Neighbors (KNN)
algorithm?
A. To ensure that the model can handle categorical data without encoding
B. To increase the likelihood of the model being linearly separable
,C. To reduce the computational complexity of calculating Euclidean distance
D. To prevent variables with larger scales from dominating the distance calculation
Answer: D
Conceptual Explanation: KNN relies on distance metrics; if one feature has a much larger
range than others, it will disproportionately influence the distance and the resulting
classification.
3. Which of the following is a significant disadvantage of the K-means clustering algorithm?
A. It is computationally too slow for large datasets
B. It cannot be used for numerical data
C. It automatically determines the optimal number of clusters
D. The results are sensitive to the initial selection of centroids
Answer: D
Conceptual Explanation: K-means is a heuristic algorithm that may converge to a local
optimum depending on where the initial centroids are placed.
4. In a CUSUM (Cumulative Sum) control chart, what does the parameter ‘H’ represent?
A. The expected mean of the process
B. The slack value to ignore small shifts
C. The standard deviation of the observations
, D. The threshold limit that triggers an alarm when exceeded
Answer: D
Conceptual Explanation: In CUSUM, H is the threshold value. If the cumulative sum
exceeds H, the process is considered to have shifted (out of control).
5. What is the primary objective of Principal Component Analysis (PCA)?
A. To maximize the correlation between the target variable and features
B. To reduce dimensionality while preserving as much variance as possible
C. To ensure that all features have a mean of zero and variance of one
D. To find the best linear regression fit for a dataset
Answer: B
Conceptual Explanation: PCA transforms a large set of variables into a smaller one that
still contains most of the original information (variance).
6. Which component of a time series represents long-term fluctuations that are not related to
calendar-based cycles?
A. Seasonality
B. Trend
C. Cyclicity
D. Random Error
