01
True or false
1 point
A retailer groups customers by purchase patterns without supplying preassigned behavior labels. Is this supervised learning because the algorithm produces group IDs?
Free Practice Quiz Question List
7 Introduction to Machine Learning Online Quiz Questions
Use this free practice quiz with 20 questions to review 7 Introduction to Machine Learning, test your knowledge, and prepare for your next test or exam.
20 questions
02
Choose one
1 point
A housing model uses home size and neighborhood to predict the sale price of homes that have not yet sold. Which task is it performing?
- A
Regression
- B
Classification
- C
Clustering
- D
Dimensionality reduction
03
Written response
1 point
When dividing classification data, what kind of split can help preserve the proportions of each class in both the training and test sets?
04
Fill in the blank
1 point
A retailer wants to estimate how a model will perform on future transactions. With records ordered by date, it should train on an earlier period and test on a later period, using a .
05
Choose one
1 point
When dividing supervised data into training and test sets, which action preserves the integrity of the examples?
- A
Keep each example's features together with its corresponding label.
- B
Split the features and labels independently into training and test sets.
- C
Remove the labels before dividing the examples.
- D
Place every example with a particular label in the test set.
06
Written response
1 point
A team uses information from the test set to decide how to fill missing values before reporting the model's test score. What is this problem called?
07
Fill in the blank
1 point
During development, use a or cross-validation to compare model choices, while reserving the test set for a final evaluation.
08
True or false
1 point
A model scores much better on its training data than on held-out data. This difference can be a warning sign of overfitting.
- A
True
- B
False
09
Choose one
1 point
A model performs poorly on both its training examples and held-out examples. Which limitation does this pattern most directly suggest?
- A
Underfitting
- B
Overfitting
- C
Data leakage
- D
Clustering
10
Choose one
1 point
A dataset contains several records for each person. To avoid an unrealistic evaluation caused by related records appearing in both sets, how should the split handle each person's records?
- A
Randomly distribute the person's records across both sets.
- B
Keep the person's related records together in one set.
- C
Put only the person's most recent record in the test set, regardless of other records.
- D
Use the person's records to preprocess the test set before splitting.
11
Choose all
1 point
Select all statements that correctly describe classification metrics.
- A
Accuracy is the fraction of predictions that are correct.
- B
Precision is the fraction of positive predictions that are correct.
- C
Recall is the fraction of actual positives that are found.
- D
Precision is the fraction of actual positives that are found.
- E
Recall is the fraction of positive predictions that are correct.
12
Choose all
1 point
A team wants to investigate whether its model may reflect bias in the data or labels. Select all useful checks supported by sound evaluation practice.
- A
Check whether relevant populations are underrepresented in the training data.
- B
Treat a strong overall score as sufficient proof that errors are even across groups.
- C
Examine whether labels reflect historically biased decisions or misleading targets.
- D
Evaluate errors for relevant subgroups.
- E
Assume that removing one metric from evaluation guarantees fairness.
13
Choose one
1 point
Two regression models have comparable typical errors, but one makes a few much larger errors. Which measure, compared with mean absolute error, gives relatively more weight to those large errors?
- A
Accuracy
- B
Root mean squared error (RMSE)
- C
Precision
- D
A clustering separation score
14
Open ended
1 point
A clustering method produces groups with a strong score for compactness and separation. Explain what that score can indicate, why it does not establish that the groups are useful in the real world, and what additional input may be needed.
15
Choose one
1 point
A model predicts the likely dollar loss from a fraudulent transaction. Which type of supervised task is this?
- A
Classification
- B
Regression
- C
Clustering
- D
Dimensionality reduction
16
Choose one
1 point
A retailer has customer purchase patterns but no target labels. It groups customers by similarity to explore different shopping behaviors. Which method best describes this task?
- A
Clustering
- B
Classification
- C
Regression
- D
Stratified splitting
17
Written response
1 point
What is the standard abbreviation for the regression metric that averages the absolute differences between predicted and actual values?
18
Choose one
1 point
A screening system should find as many actual positive cases as possible, even if it produces some false alarms. Which classification metric most directly reflects that priority?
- A
Accuracy
- B
Precision
- C
Recall
- D
Mean absolute error
19
True or false
1 point
A model scores highly on a held-out test set, so a team concludes it is guaranteed to perform equally well in every real-world setting. Is that conclusion correct?
- A
True
- B
False
20
Written response
1 point
A model performs poorly even on its training data because it fails to capture useful patterns. What is this problem called?