10.8.1: Review Questions
- Page ID
- 58335
\( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \)
\( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)
\( \newcommand{\dsum}{\displaystyle\sum\limits} \)
\( \newcommand{\dint}{\displaystyle\int\limits} \)
\( \newcommand{\dlim}{\displaystyle\lim\limits} \)
\( \newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\)
( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\)
\( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\)
\( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\)
\( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\)
\( \newcommand{\Span}{\mathrm{span}}\)
\( \newcommand{\id}{\mathrm{id}}\)
\( \newcommand{\Span}{\mathrm{span}}\)
\( \newcommand{\kernel}{\mathrm{null}\,}\)
\( \newcommand{\range}{\mathrm{range}\,}\)
\( \newcommand{\RealPart}{\mathrm{Re}}\)
\( \newcommand{\ImaginaryPart}{\mathrm{Im}}\)
\( \newcommand{\Argument}{\mathrm{Arg}}\)
\( \newcommand{\norm}[1]{\| #1 \|}\)
\( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\)
\( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\AA}{\unicode[.8,0]{x212B}}\)
\( \newcommand{\vectorA}[1]{\vec{#1}} % arrow\)
\( \newcommand{\vectorAt}[1]{\vec{\text{#1}}} % arrow\)
\( \newcommand{\vectorB}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \)
\( \newcommand{\vectorC}[1]{\textbf{#1}} \)
\( \newcommand{\vectorD}[1]{\overrightarrow{#1}} \)
\( \newcommand{\vectorDt}[1]{\overrightarrow{\text{#1}}} \)
\( \newcommand{\vectE}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash{\mathbf {#1}}}} \)
\( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \)
\(\newcommand{\longvect}{\overrightarrow}\)
\( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)
\(\newcommand{\ket}[1]{\left| #1 \right>}\)
\(\newcommand{\bra}[1]{\left< #1 \right|}\)
\(\newcommand{\braket}[2]{\left< #1 \vphantom{#2} \right| \left. #2 \vphantom{#1} \right>}\)
\(\newcommand{\braopket}[3]{\left< #1 \vphantom{#2}\vphantom{#3} \right| #2 \vphantom{#1}\vphantom{#3} \left| #3 \vphantom{#1}\vphantom{#2} \right>}\)
\(\newcommand{\qmvec}[1]{\mathbf{\vec{#1}}}\)
\(\newcommand{\op}[1]{\hat{\mathbf{#1}}}\)
\(\newcommand{\expect}[1]{\langle #1 \rangle}\)
\(\newcommand{\dfn}[1]{\emph{\textbf{#1}}}\)
Review Questions
Multiple Choice
-
Predicting whether a customer will churn (yes/no) is an example of:
-
a) Regression b) Classification c) Clustering d) Dimensionality reduction
-
-
A model achieves 99% accuracy on training data but only 60% on test data. This is a sign of:
-
a) Underfitting b) Overfitting c) Good generalization d) High bias
-
-
K-means clustering is an example of:
-
a) Supervised classification b) Supervised regression c) Unsupervised learning d) Semi-supervised learning
-
Short Answer
-
Explain the difference between supervised and unsupervised learning. Give one real-world example of each in a business context.
-
Compare linear regression and logistic regression. What type of problem does each solve? What is the key mathematical difference (continuous output vs. probability)?
-
What is overfitting, and why is it a problem? Describe two strategies to prevent it.
-
Explain precision vs. recall using a spam email filter example. In this context, which metric is more important, and why?
True/False
-
True or False: A decision tree can be used for both classification and regression tasks.
-
True or False: K-means clustering requires you to specify the number of clusters (K) in advance.
-
True or False: Higher model accuracy always means the model is better for the business problem.
Answer Key:
1-b, 2-b, 3-c, 4. Supervised: model trained with labels (e.g., predicting loan default from historical labeled data). Unsupervised: no labels (e.g., customer segmentation by purchasing behavior to discover natural groups). 5. Linear regression: predicts continuous numeric output (e.g., house price). Logistic regression: predicts probability of a binary class (e.g., churn yes/no). Linear outputs unbounded numbers; logistic applies sigmoid to squash output to 0-1 probability. 6. Overfitting: model memorizes training data noise, fails on new data. Strategies: (1) train/test split or cross-validation to detect it; (2) limit model complexity (prune trees, add regularization); (3) gather more training data. 7. Precision: of emails flagged as spam, how many truly are (if low, legitimate emails go to spam — annoying). Recall: of all actual spam, how many are caught (if low, spam reaches inbox). For spam filters, precision is often prioritized — users hate losing important emails more than receiving occasional spam. 8. True, 9. True, 10. False (e.g., 95% accuracy on imbalanced data where 95% of cases are one class is meaningless — need precision/recall).