ISYE 6501 – INTRODUCTION TO ANALYTICS MODELING
COMPLETE EXAM PREPARATION GUIDE 200 QUESTIONS WITH
VERIFIED ANSWERS AND DETAILED RATIONALES GEORGIA TECH / EDX
COURSE – LATEST UPDATE
SECTION 1: FOUNDATIONAL CONCEPTS & MODEL TYPES (Questions 1–30)
Question 1
Which of the following best defines "analytics modeling"?
A) The process of storing large amounts of data
B) The process of using mathematical and statistical techniques to analyze data
and make decisions
C) The process of collecting data from multiple sources
D) The process of visualizing data in dashboards
Answer: B) The process of using mathematical and statistical techniques to
analyze data and make decisions
Rationale: Analytics modeling is the core of data science where mathematical,
statistical, and computational methods are applied to data. Its purpose is to
extract meaningful insights, detect patterns, and support evidence-based
decision-making. While data collection, storage, and visualization are
important steps in the broader analytics pipeline, modeling specifically
refers to the analytical techniques that transform raw data into actionable
knowledge.
Question 2
What do descriptive questions in analytics ask?
A) What will happen?
B) What action(s) would be best?
C) What happened?
D) Why did it happen?
Answer: C) What happened?
Rationale: Descriptive analytics is the most basic form of analytics and
focuses on summarizing historical data to understand what has occurred.
This includes generating reports, dashboards, and aggregations. Predictive
1
,analytics addresses "what will happen?"; prescriptive analytics asks "what
should we do?"; and diagnostic analytics explores "why did it happen?"
Each type builds upon the previous one to provide deeper insights.
Question 3
Which of the following is an example of a supervised learning problem?
A) Grouping customers into segments based on purchasing behavior
B) Predicting whether a customer will churn based on historical data
C) Finding patterns in transaction data without predefined labels
D) Reducing the dimensionality of a dataset
Answer: B) Predicting whether a customer will churn based on historical data
Rationale: Supervised learning requires labeled data where the outcome
(target variable) is known. Predicting customer churn uses historical
data with known outcomes (churned or not), making it a classification
problem. Grouping customers, finding patterns without labels, and
dimensionality reduction are all unsupervised learning tasks because
they do not use a predefined target variable.
Question 4
In the context of linear regression, which of the following is a consequence
of multicollinearity?
A) The model becomes more accurate
B) The coefficients become unstable and difficult to interpret
C) The R-squared value decreases significantly
D) The model cannot be estimated
Answer: B) The coefficients become unstable and difficult to interpret
Rationale: Multicollinearity occurs when two or more predictor variables
are highly correlated with each other. This causes the coefficient
estimates to become very sensitive to small changes in the data,
inflating their standard errors and making it difficult to determine
the individual effect of each predictor. The model can still be
estimated, but the interpretation of individual coefficients becomes
2
,unreliable, and the model's overall predictive power is not necessarily
reduced.
Question 5
What does the p-value for a coefficient in a linear regression model indicate?
A) The probability that the coefficient is exactly zero
B) The probability of observing the coefficient (or more extreme) if the null
hypothesis is true
C) The magnitude of the coefficient's effect
D) The confidence interval for the coefficient
Answer: B) The probability of observing the coefficient (or more extreme) if
the null hypothesis is true
Rationale: The p-value is a measure of statistical significance. It tests
the null hypothesis that the true coefficient is zero, meaning the
predictor has no effect on the response. A low p-value (typically below
0.05) suggests that the observed relationship is unlikely to have occurred
by random chance, providing evidence that the predictor has a real effect.
The p-value does not indicate the size or importance of the effect.
Question 6
What is the primary purpose of cross-validation in analytics modeling?
A) To test the model on the training data
B) To evaluate how well the model will generalize to new data
C) To increase the complexity of the model
D) To reduce the number of features
Answer: B) To evaluate how well the model will generalize to new data
Rationale: Cross-validation is a resampling technique used to assess a
model's performance on unseen data. It splits the dataset into multiple
folds, trains the model on a subset, and validates it on the remaining
fold, repeating this process several times. This provides a more robust
estimate of generalization error than using a single train-test split
and helps detect overfitting, where a model performs well on training
3
, data but poorly on new data.
Question 7
Which of the following best describes "overfitting" in a model?
A) The model performs well on training data but poorly on new data
B) The model performs poorly on both training and new data
C) The model is too simple to capture the underlying patterns
D) The model has high bias
Answer: A) The model performs well on training data but poorly on new data
Rationale: Overfitting occurs when a model learns the noise and random
fluctuations in the training data instead of the true underlying pattern.
This results in a model that fits the training data extremely well but
fails to generalize to new, unseen data. Overfitting is often caused by
using a model that is too complex for the amount of training data
available, such as a high-degree polynomial or a very deep decision tree.
Question 8
What types of models typically require standardized data?
A) Tree-based models
B) Distance-based models (e.g., k-nearest neighbors, clustering, SVM)
C) Linear regression models
D) Time series models
Answer: B) Distance-based models (e.g., k-nearest neighbors, clustering, SVM)
Rationale: Distance-based models compute distances between data points to
determine similarity or proximity. If features are on different scales
(age in years vs. income in dollars), features with larger magnitudes
will dominate the distance calculation. Standardization (z-score) or
min-max scaling ensures that all features contribute equally, preventing
bias toward features with larger numerical ranges and improving model
performance.
4
COMPLETE EXAM PREPARATION GUIDE 200 QUESTIONS WITH
VERIFIED ANSWERS AND DETAILED RATIONALES GEORGIA TECH / EDX
COURSE – LATEST UPDATE
SECTION 1: FOUNDATIONAL CONCEPTS & MODEL TYPES (Questions 1–30)
Question 1
Which of the following best defines "analytics modeling"?
A) The process of storing large amounts of data
B) The process of using mathematical and statistical techniques to analyze data
and make decisions
C) The process of collecting data from multiple sources
D) The process of visualizing data in dashboards
Answer: B) The process of using mathematical and statistical techniques to
analyze data and make decisions
Rationale: Analytics modeling is the core of data science where mathematical,
statistical, and computational methods are applied to data. Its purpose is to
extract meaningful insights, detect patterns, and support evidence-based
decision-making. While data collection, storage, and visualization are
important steps in the broader analytics pipeline, modeling specifically
refers to the analytical techniques that transform raw data into actionable
knowledge.
Question 2
What do descriptive questions in analytics ask?
A) What will happen?
B) What action(s) would be best?
C) What happened?
D) Why did it happen?
Answer: C) What happened?
Rationale: Descriptive analytics is the most basic form of analytics and
focuses on summarizing historical data to understand what has occurred.
This includes generating reports, dashboards, and aggregations. Predictive
1
,analytics addresses "what will happen?"; prescriptive analytics asks "what
should we do?"; and diagnostic analytics explores "why did it happen?"
Each type builds upon the previous one to provide deeper insights.
Question 3
Which of the following is an example of a supervised learning problem?
A) Grouping customers into segments based on purchasing behavior
B) Predicting whether a customer will churn based on historical data
C) Finding patterns in transaction data without predefined labels
D) Reducing the dimensionality of a dataset
Answer: B) Predicting whether a customer will churn based on historical data
Rationale: Supervised learning requires labeled data where the outcome
(target variable) is known. Predicting customer churn uses historical
data with known outcomes (churned or not), making it a classification
problem. Grouping customers, finding patterns without labels, and
dimensionality reduction are all unsupervised learning tasks because
they do not use a predefined target variable.
Question 4
In the context of linear regression, which of the following is a consequence
of multicollinearity?
A) The model becomes more accurate
B) The coefficients become unstable and difficult to interpret
C) The R-squared value decreases significantly
D) The model cannot be estimated
Answer: B) The coefficients become unstable and difficult to interpret
Rationale: Multicollinearity occurs when two or more predictor variables
are highly correlated with each other. This causes the coefficient
estimates to become very sensitive to small changes in the data,
inflating their standard errors and making it difficult to determine
the individual effect of each predictor. The model can still be
estimated, but the interpretation of individual coefficients becomes
2
,unreliable, and the model's overall predictive power is not necessarily
reduced.
Question 5
What does the p-value for a coefficient in a linear regression model indicate?
A) The probability that the coefficient is exactly zero
B) The probability of observing the coefficient (or more extreme) if the null
hypothesis is true
C) The magnitude of the coefficient's effect
D) The confidence interval for the coefficient
Answer: B) The probability of observing the coefficient (or more extreme) if
the null hypothesis is true
Rationale: The p-value is a measure of statistical significance. It tests
the null hypothesis that the true coefficient is zero, meaning the
predictor has no effect on the response. A low p-value (typically below
0.05) suggests that the observed relationship is unlikely to have occurred
by random chance, providing evidence that the predictor has a real effect.
The p-value does not indicate the size or importance of the effect.
Question 6
What is the primary purpose of cross-validation in analytics modeling?
A) To test the model on the training data
B) To evaluate how well the model will generalize to new data
C) To increase the complexity of the model
D) To reduce the number of features
Answer: B) To evaluate how well the model will generalize to new data
Rationale: Cross-validation is a resampling technique used to assess a
model's performance on unseen data. It splits the dataset into multiple
folds, trains the model on a subset, and validates it on the remaining
fold, repeating this process several times. This provides a more robust
estimate of generalization error than using a single train-test split
and helps detect overfitting, where a model performs well on training
3
, data but poorly on new data.
Question 7
Which of the following best describes "overfitting" in a model?
A) The model performs well on training data but poorly on new data
B) The model performs poorly on both training and new data
C) The model is too simple to capture the underlying patterns
D) The model has high bias
Answer: A) The model performs well on training data but poorly on new data
Rationale: Overfitting occurs when a model learns the noise and random
fluctuations in the training data instead of the true underlying pattern.
This results in a model that fits the training data extremely well but
fails to generalize to new, unseen data. Overfitting is often caused by
using a model that is too complex for the amount of training data
available, such as a high-degree polynomial or a very deep decision tree.
Question 8
What types of models typically require standardized data?
A) Tree-based models
B) Distance-based models (e.g., k-nearest neighbors, clustering, SVM)
C) Linear regression models
D) Time series models
Answer: B) Distance-based models (e.g., k-nearest neighbors, clustering, SVM)
Rationale: Distance-based models compute distances between data points to
determine similarity or proximity. If features are on different scales
(age in years vs. income in dollars), features with larger magnitudes
will dominate the distance calculation. Standardization (z-score) or
min-max scaling ensures that all features contribute equally, preventing
bias toward features with larger numerical ranges and improving model
performance.
4
