ISYE 6501 Final Exam Quiz Bank: Over 56
Questions with Detailed Explanations | ISYE
6501 Study Guide | Georgia Institute of
Technology
This expanded set covers the major topics in ISYE 6501: model evaluation,
regression (linear/logistic/regularization), classification (KNN/SVM/trees),
clustering (K-means/hierarchical), PCA, time series (ARIMA/GARCH/smoothing),
optimization (explore/exploit, A/B testing), and conceptual best practices. Good
luck on your final quiz!
Topic 1: Model Evaluation & Basic Concepts
Q1: What does R² measure?
A. Error rate
B. Model complexity
C. Variance explained
D. Bias
Answer: C. Variance explained
Rationale: R² indicates the proportion of the variance in the dependent variable
that is explained by the independent variables. It ranges from 0 to 1 .
Q2: What is overfitting?
A. Underestimating error
B. Model too simple
C. Model fits noise
, D. Data cleaning
Answer: C. Model fits noise
Rationale: Overfitting occurs when a model captures random noise and
fluctuations in the training data instead of the true underlying signal, leading to
poor generalization on new data .
Q3: Which method reduces overfitting?
A. Increase variables
B. Cross-validation
C. Ignore data
D. Random guessing
Answer: B. Cross-validation
Rationale: Cross-validation evaluates model generalization by training on multiple
subsets and testing on held-out data, helping detect when a model fails to
generalize .
Q4: What does RMSE stand for?
A. Mean error
B. Root mean squared error
C. Regression score
D. Residual sum
Answer: B. Root mean squared error
Rationale: RMSE measures the average magnitude of prediction errors by taking
the square root of the average squared differences between predicted and actual
values .
Q5: What is the F1 score?
A. Accuracy & recall
B. Precision & recall
C. Bias & variance
, D. Error & loss
Answer: B. Precision & recall
Rationale: The F1 score is the harmonic mean of precision and recall, providing a
balanced metric for classification performance, especially on imbalanced
datasets .
Topic 2: Regression (Linear, Logistic, Regularization)
Q6: In logistic regression, the output of the model is a probability between 0
and 1. If you lower the classification threshold to 0.1, what is the expected
outcome?
A. You will catch more true positives, but also incorrectly flag many
negatives as positives
B. You will miss almost all positives
C. The model will switch to a decision tree format
D. You will perfectly balance false positives and false negatives
Answer: A
Rationale: Lowering the threshold makes the model more sensitive. It will catch
more true positives but at the cost of increasing false positives .
Q7: What is the primary purpose of LASSO regression (L1 regularization)?
A. To increase model complexity
B. To perform variable selection by shrinking some coefficients to zero
C. To combine multiple trees
D. To cluster data points
Answer: B
Rationale: LASSO (Least Absolute Shrinkage and Selection Operator) adds a
penalty equal to the sum of the absolute values of coefficients. This can force
some coefficients to become exactly zero, effectively performing variable
selection .
Questions with Detailed Explanations | ISYE
6501 Study Guide | Georgia Institute of
Technology
This expanded set covers the major topics in ISYE 6501: model evaluation,
regression (linear/logistic/regularization), classification (KNN/SVM/trees),
clustering (K-means/hierarchical), PCA, time series (ARIMA/GARCH/smoothing),
optimization (explore/exploit, A/B testing), and conceptual best practices. Good
luck on your final quiz!
Topic 1: Model Evaluation & Basic Concepts
Q1: What does R² measure?
A. Error rate
B. Model complexity
C. Variance explained
D. Bias
Answer: C. Variance explained
Rationale: R² indicates the proportion of the variance in the dependent variable
that is explained by the independent variables. It ranges from 0 to 1 .
Q2: What is overfitting?
A. Underestimating error
B. Model too simple
C. Model fits noise
, D. Data cleaning
Answer: C. Model fits noise
Rationale: Overfitting occurs when a model captures random noise and
fluctuations in the training data instead of the true underlying signal, leading to
poor generalization on new data .
Q3: Which method reduces overfitting?
A. Increase variables
B. Cross-validation
C. Ignore data
D. Random guessing
Answer: B. Cross-validation
Rationale: Cross-validation evaluates model generalization by training on multiple
subsets and testing on held-out data, helping detect when a model fails to
generalize .
Q4: What does RMSE stand for?
A. Mean error
B. Root mean squared error
C. Regression score
D. Residual sum
Answer: B. Root mean squared error
Rationale: RMSE measures the average magnitude of prediction errors by taking
the square root of the average squared differences between predicted and actual
values .
Q5: What is the F1 score?
A. Accuracy & recall
B. Precision & recall
C. Bias & variance
, D. Error & loss
Answer: B. Precision & recall
Rationale: The F1 score is the harmonic mean of precision and recall, providing a
balanced metric for classification performance, especially on imbalanced
datasets .
Topic 2: Regression (Linear, Logistic, Regularization)
Q6: In logistic regression, the output of the model is a probability between 0
and 1. If you lower the classification threshold to 0.1, what is the expected
outcome?
A. You will catch more true positives, but also incorrectly flag many
negatives as positives
B. You will miss almost all positives
C. The model will switch to a decision tree format
D. You will perfectly balance false positives and false negatives
Answer: A
Rationale: Lowering the threshold makes the model more sensitive. It will catch
more true positives but at the cost of increasing false positives .
Q7: What is the primary purpose of LASSO regression (L1 regularization)?
A. To increase model complexity
B. To perform variable selection by shrinking some coefficients to zero
C. To combine multiple trees
D. To cluster data points
Answer: B
Rationale: LASSO (Least Absolute Shrinkage and Selection Operator) adds a
penalty equal to the sum of the absolute values of coefficients. This can force
some coefficients to become exactly zero, effectively performing variable
selection .
