ISYE 6501 Final Exam Comprehensive Study Guide and Practice Test Bank for 2026/2027 Academic Year

2026/2027


ISYE 6501 Final Exam Comprehensive
Study Guide and Practice Test Bank for
2026/2027 Academic Year

Question 1:

What type of problem is linear regression best suited for?

A. Classification of categorical outcomes
B. Prediction of continuous numerical values
C. Grouping unlabeled data into clusters
D. Experimental design optimization

Correct Answer: B. Prediction of continuous numerical values

Rationale: Linear regression is used when the goal is to predict a continuous outcome
(such as sales, temperature, or price) based on one or more predictor variables. It
models a linear relationship between inputs and outputs. It is not used for
classification, clustering, or experimental design tasks.


Question 2:

What type of problem is fractional factorial design best suited for?

A. Time-series forecasting
B. Variable selection in regression
C. Efficient experimental design with many factors
D. Classification of binary outcomes

Correct Answer: C. Efficient experimental design with many factors

Rationale: Fractional factorial design is used in experimental settings where testing all
combinations of factors is too costly. It efficiently selects a subset of experimental
runs while still capturing key factor effects. It is not used for prediction or
classification tasks.


Question 3:

What type of problem is lasso regression best suited for?

A. Clustering unlabeled data
B. Variable selection and prediction
C. Time-series volatility forecasting
D. Image classification

,2026/2027

Correct Answer: B. Variable selection and prediction

Rationale: Lasso regression is a regression method that performs both prediction and
automatic variable selection by shrinking less important coefficients to zero. This
helps improve model interpretability and reduce overfitting.


Question 4:

What type of problem is logistic regression best suited for?

A. Predicting continuous outcomes
B. Clustering similar observations
C. Classification and probability estimation
D. Designing experiments

Correct Answer: C. Classification and probability estimation

Rationale: Logistic regression is used for classification problems where the output is
categorical (often binary). It estimates probabilities that an observation belongs to a
class. It is not used for clustering or continuous prediction.


Question 5:

What type of problem is k-means clustering best suited for?

A. Supervised classification
B. Regression modeling
C. Unsupervised clustering
D. Time-series forecasting

Correct Answer: C. Unsupervised clustering

Rationale: K-means clustering groups unlabeled data into clusters based on similarity.
It is an unsupervised learning method and does not rely on known outcomes or labels.


Question 6:

What type of problem is a support vector machine (SVM) best suited for?

A. Predicting continuous values
B. Classification problems with feature data
C. Time-series forecasting
D. Experimental design

Correct Answer: B. Classification problems with feature data

,2026/2027

Rationale: Support vector machines are supervised learning models primarily used for
classification. They find an optimal boundary (hyperplane) that separates classes in
feature space.


Question 7:

What type of problem is GARCH best suited for?

A. Predicting categorical outcomes
B. Forecasting time-series volatility
C. Clustering observations
D. Designing experiments

Correct Answer: B. Forecasting time-series volatility

Rationale: GARCH models are used in time-series analysis to predict changing
variance (volatility), especially in financial data. They are not used for classification
or clustering.


Question 8:

What type of problem is factorial design best suited for?

A. Predicting binary outcomes
B. Efficient experimental testing of all factor combinations
C. Forecasting time-series trends
D. Clustering data points

Correct Answer: B. Efficient experimental testing of all factor combinations

Rationale: Factorial design tests all combinations of factors in an experiment to
understand their effects. It is used in experimental design, not prediction or clustering.


Question 9:

Linear regression is best described as:

A. Predicting whether something happens
B. Predicting an amount using feature data
C. Grouping similar observations
D. Forecasting volatility

Correct Answer: B. Predicting an amount using feature data

Rationale: Linear regression uses feature variables to predict a continuous numerical
outcome. It is commonly used for forecasting quantities such as sales or demand.

, 2026/2027


Question 10:

K-nearest neighbors classification is best described as:

A. Predicting a continuous amount
B. Predicting whether something will happen
C. Clustering unlabeled data
D. Designing experiments

Correct Answer: B. Predicting whether something will happen

Rationale: K-nearest neighbors classification predicts class membership (yes/no or
category) based on similarity to nearby data points in feature space.


Question 11:

GARCH is best described as:

A. Predicting variance using time-series data
B. Predicting categorical outcomes
C. Grouping similar observations
D. Designing experiments

Correct Answer: A. Predicting variance using time-series data

Rationale: GARCH models are designed to estimate volatility (variance) over time in
time-series data, especially in finance.


Question 12:

ARIMA is best described as:

A. Predicting categorical outcomes
B. Predicting amounts using time-series data
C. Clustering observations
D. Experimental design

Correct Answer: B. Predicting amounts using time-series data

Rationale: ARIMA is a time-series forecasting method used to predict future
numerical values based on past trends and patterns.


Question 13:

Is the following time-series data? “Number of burgers sold each day for 2000 days”