ISYE 6501 Final Exam Questions and Answers Latest Update This Year

ISYE 6501 Final Exam Questions and
Answers Latest Update This Year

SECTION 1: MODEL SELECTION & PROBLEM TYPES
Identifying the correct analytical tool for a specific business or statistical
problem is a core focus of the ISYE 6501 final exam .


Question 1: For which type of problem is linear regression best suited?
A) Classification
B) Clustering
C) Experimental design
D) Prediction from feature data
E) Prediction from time-series data
F) Variable selection
Answer: D – Prediction from feature data
Explanation: Linear regression models the relationship between a continuous
dependent variable and one or more independent (feature) variables. It is used
for predicting numerical outcomes based on feature data, such as predicting
house prices from square footage and number of bedrooms .


Question 2: For which type of problem is logistic regression best suited?
A) Classification
B) Clustering
C) Experimental design
D) Prediction from feature data
E) Prediction from time-series data
F) Variable selection
Answer: A/D – Classification and/or prediction from feature data
Explanation: Logistic regression models the probability of a binary outcome
using a logit link function, making it ideal for binary classification problems. It
predicts probabilities between 0 and 1, which can be thresholded to create
classifications .


Question 3: For which type of problem is k-means clustering best suited?



1

,A) Clustering
B) Classification
C) Experimental design
D) Prediction from feature data
E) Prediction from time-series data
F) Variable selection
Answer: A – Clustering
Explanation: K-means is an unsupervised learning algorithm that partitions data
into k distinct clusters based on feature similarity. It does not use labeled data,
distinguishing it from classification methods .


Question 4: For which type of problem is Principal Components Analysis (PCA)
best suited?
A) Classification
B) Clustering
C) Experimental design
D) Prediction from feature data
E) Prediction from time-series data
F) Variable selection / feature extraction
Answer: F – Variable selection / feature extraction
Explanation: PCA reduces the dimensionality of a dataset by creating new
uncorrelated components that capture maximum variance. It is a feature
extraction technique that transforms original variables into principal
components .


Question 5: For which type of problem is ARIMA best suited?
A) Classification
B) Clustering
C) Experimental design
D) Prediction from feature data
E) Prediction from time-series data
F) Variable selection
Answer: E – Prediction from time-series data
Explanation: ARIMA (AutoRegressive Integrated Moving Average) is a class of
models designed specifically for analyzing and forecasting time-series data.
The 'I' stands for differencing to achieve stationarity .


Question 6: For which type of problem is GARCH best suited?

2

,A) Classification
B) Clustering
C) Experimental design
D) Prediction from feature data
E) Prediction from time-series data
F) Variable selection
Answer: E – Prediction from time-series data
Explanation: GARCH (Generalized Autoregressive Conditional
Heteroskedasticity) models are specifically designed to model and forecast
changing volatility in time-series data, particularly in financial applications like
stock market volatility .


Question 7: For which type of problem is k-nearest neighbors (K-NN) best
suited?
A) Classification
B) Clustering
C) Experimental design
D) Prediction from feature data
E) Prediction from time-series data
F) Variable selection
Answer: A/D – Classification and/or prediction from feature data
Explanation: K-NN is a non-parametric method used for both classification
(discrete outcomes) and regression (continuous prediction) based on feature
data. It works by finding the k closest training examples in the feature space .


Question 8: For which type of problem is CART (Classification and Regression
Trees) best suited?
A) Clustering
B) Classification and/or regression (prediction from feature data)
C) Experimental design
D) Prediction from time-series data
E) Variable selection
Answer: B – Classification and/or regression (prediction from feature data)
Explanation: CART builds decision trees for both classification and regression
tasks. It recursively partitions the feature space to make predictions .


Question 9: For which type of problem is exponential smoothing best suited?



3

, A) Classification
B) Clustering
C) Variable selection
D) Experimental design
E) Prediction from time-series data
Answer: E – Prediction from time-series data
Explanation: Exponential smoothing is a time-series forecasting method that
assigns exponentially decreasing weights to older observations. It has three
variants: single (no trend/seasonality), double (trend), and triple (trend +
seasonality/Holt-Winters) .


Question 10: For which type of problem is fractional factorial design best
suited?
A) Clustering
B) Classification
C) Prediction from time-series data
D) Experimental design
E) Variable selection and/or prediction from feature data
Answer: D – Experimental design
Explanation: Fractional factorial designs are used to study the effects of
multiple factors efficiently by testing only a subset of all possible factor
combinations, making them ideal for screening experiments to identify which
factors matter most .


Question 11: For which type of problem is lasso regression best suited?
A) Clustering
B) Classification
C) Experimental design
D) Prediction from feature data
E) Prediction from time-series data
F) Variable selection
Answer: F/D – Variable selection and/or prediction from feature data
Explanation: Lasso regression (L1 regularization) shrinks some regression
coefficients to exactly zero, effectively selecting a subset of relevant features. It
is particularly useful when the number of predictors is large .


Question 12: What is the elbow method used to determine?



4