Supervised vs Unsupervised Learning Exam Question Explained
This AWS AI Practitioner question checks whether you can identify the learning type from the data available. The important clue is that the customer support tickets have no category labels.
Short answer
The correct answer is A. Unsupervised learning with clustering.
Unsupervised learning with clustering is correct because the dataset has no category labels and the team wants to discover natural groupings. In supervised learning, the model learns from examples that already include the correct answer, such as ticket text paired with a known support category. This scenario explicitly says there are no category labels, so supervised classification does not have the training signal it needs. Clustering is the common unsupervised technique for finding groups in unlabeled data. It can reveal that tickets naturally cluster around billing, login problems, device setup, refunds, or other themes even when nobody has labeled those themes ahead of time. The purpose also matches unsupervised learning: the team is exploring structure so it can design a routing system. On the exam, words like "unlabeled," "discover groups," "find patterns," "segment customers," or "cluster documents" should make you think unsupervised learning first.
The Question
A data science team has a dataset of 50,000 customer support tickets, none of which have category labels. They want to discover natural groupings of tickets to design a new routing system. Which type of machine learning is the best fit?
Why A is correct
Unsupervised learning with clustering is correct because the dataset has no category labels and the team wants to discover natural groupings. In supervised learning, the model learns from examples that already include the correct answer, such as ticket text paired with a known support category. This scenario explicitly says there are no category labels, so supervised classification does not have the training signal it needs. Clustering is the common unsupervised technique for finding groups in unlabeled data. It can reveal that tickets naturally cluster around billing, login problems, device setup, refunds, or other themes even when nobody has labeled those themes ahead of time. The purpose also matches unsupervised learning: the team is exploring structure so it can design a routing system. On the exam, words like "unlabeled," "discover groups," "find patterns," "segment customers," or "cluster documents" should make you think unsupervised learning first.
Why the other options are wrong
Supervised classification requires labeled training examples. If each ticket already had a known category, classification would be reasonable. The question says none of the tickets have labels.
Reinforcement learning is for an agent learning actions through rewards and penalties. It is not the right fit for grouping a static dataset of support tickets.
Semi-supervised learning uses a small labeled set plus a larger unlabeled set. The question gives no labeled anchors, so plain unsupervised clustering is the cleaner answer.
The exam rule for supervised vs unsupervised learning
For AIF-C01, classify the machine learning type by asking what feedback signal is available. Supervised learning needs labeled examples: inputs paired with correct outputs. A model might learn to classify emails as spam or not spam because the training data already marks which emails were spam. Regression is also supervised when the label is a number, such as predicted sales or house price. Unsupervised learning does not receive correct answers. It looks for hidden structure in the data, such as clusters, segments, anomalies, or lower-dimensional patterns. If a question says the data is unlabeled and the goal is to discover groups, unsupervised learning is usually the answer. Reinforcement learning is different again: an agent takes actions in an environment and learns from rewards. Semi-supervised learning sits between supervised and unsupervised learning and requires at least some labeled data. Exam questions often bury the clue in business wording. "We do not know the categories yet" means unsupervised. "We have historical examples with the correct class" means supervised classification. "We have historical examples with a numeric target" means supervised regression.
Ready to see how you'd score?
Take the free practice quiz and find out which AWS Certified AI Practitioner domains you need to focus on. No signup required.
Practice 5 similar questions
Same cert, same or adjacent domain. Use these after reviewing the explanation.