ISYE 6501 Final Quizzes V1, V2 & V3 | Questions and Answers |
Latest 2025/26 Update | 100% Correct – Georgia Institute of
Technology
Version 1 – Questions 1–84
1. In a linear regression model, what does the p-value for a
coefficient indicate?
A) The magnitude of the coefficient's effect
B) The probability of observing the coefficient (or more extreme) if
the true coefficient is zero
C) The confidence interval for the coefficient
D) The correlation between the predictor and response
Answer B: The probability of observing the coefficient (or more
extreme) if the true coefficient is zero
,Rationale: A low p-value (<0.05) suggests the predictor is statistically
significant.
2. Which validation method is most computationally efficient for
large datasets?
A) k-fold cross-validation
B) Leave-one-out cross-validation
C) Validation set approach (holdout)
D) Bootstrap
Answer C: Validation set approach (holdout)
Rationale: Uses a single train-test split; less computation than cross-
validation.
3. In k-nearest neighbors (k-NN), what happens if k is set to 1?
A) The model underfits
B) The model overfits; decision boundary is highly irregular
C) The model is perfectly generalizable
D) The model ignores all training data
Answer B: The model overfits; decision boundary is highly irregular
,Rationale: Low k captures noise, leading to high variance and
overfitting.
4. Which clustering algorithm requires the number of clusters as an
input?
A) Hierarchical clustering
B) k-means
C) DBSCAN
D) Gaussian mixture models
Answer B: k-means
Rationale: k-means needs k specified; hierarchical clustering does
not.
5. LASSO regression can shrink coefficients exactly to zero. True or
False?
A) False
B) True
C) Only with interaction terms
D) Only with standardized data
, Answer B: True
Rationale: The L1 penalty used in LASSO produces sparse solutions.
6. Which model is best for binary classification with a linear
boundary?
A) Support vector machine with RBF kernel
B) Logistic regression
C) k-nearest neighbors
D) Random forest
Answer B: Logistic regression
Rationale: Logistic regression outputs probabilities and works well for
binary outcomes.
7. Overfitting is more likely when model complexity is ______.
A) Low
B) High
C) Moderate
D) Irrelevant
Answer B: High
Latest 2025/26 Update | 100% Correct – Georgia Institute of
Technology
Version 1 – Questions 1–84
1. In a linear regression model, what does the p-value for a
coefficient indicate?
A) The magnitude of the coefficient's effect
B) The probability of observing the coefficient (or more extreme) if
the true coefficient is zero
C) The confidence interval for the coefficient
D) The correlation between the predictor and response
Answer B: The probability of observing the coefficient (or more
extreme) if the true coefficient is zero
,Rationale: A low p-value (<0.05) suggests the predictor is statistically
significant.
2. Which validation method is most computationally efficient for
large datasets?
A) k-fold cross-validation
B) Leave-one-out cross-validation
C) Validation set approach (holdout)
D) Bootstrap
Answer C: Validation set approach (holdout)
Rationale: Uses a single train-test split; less computation than cross-
validation.
3. In k-nearest neighbors (k-NN), what happens if k is set to 1?
A) The model underfits
B) The model overfits; decision boundary is highly irregular
C) The model is perfectly generalizable
D) The model ignores all training data
Answer B: The model overfits; decision boundary is highly irregular
,Rationale: Low k captures noise, leading to high variance and
overfitting.
4. Which clustering algorithm requires the number of clusters as an
input?
A) Hierarchical clustering
B) k-means
C) DBSCAN
D) Gaussian mixture models
Answer B: k-means
Rationale: k-means needs k specified; hierarchical clustering does
not.
5. LASSO regression can shrink coefficients exactly to zero. True or
False?
A) False
B) True
C) Only with interaction terms
D) Only with standardized data
, Answer B: True
Rationale: The L1 penalty used in LASSO produces sparse solutions.
6. Which model is best for binary classification with a linear
boundary?
A) Support vector machine with RBF kernel
B) Logistic regression
C) k-nearest neighbors
D) Random forest
Answer B: Logistic regression
Rationale: Logistic regression outputs probabilities and works well for
binary outcomes.
7. Overfitting is more likely when model complexity is ______.
A) Low
B) High
C) Moderate
D) Irrelevant
Answer B: High
