Question 35
Content Domain 3: ModelingA data team has a historical dataset in which each customer record includes whether the customer churned. They want to train a model to predict churn for future customers. Which type of machine learning is most appropriate for this task?
Correct answer: A
Explanation
Use supervised learning when the dataset includes known target outcomes to predict; use unsupervised learning when the goal is to find structure in data without labeled outcomes. — Know the difference between supervised and unsupervised learning.
Why each option is right or wrong
A. Supervised learning
The dataset includes a known outcome variable—whether each customer churned—which serves as the label for training a predictive model. That makes this a supervised learning problem because the model learns from input records paired with correct outputs.
B. Unsupervised learning
Unsupervised learning applies when data has no labeled target outcome to predict.
C. Unsupervised learning, because churn is a future event
Prediction timing does not determine the category; the presence of labeled outcomes does.
D. Either supervised or unsupervised learning, since both use customer records
The distinction depends on whether labeled outcomes are available, not on using the same type of records.