Courseiva

AI-900 Practice Question: Describe fundamental principles of machine learning on Azure

A retail company wants to analyze customer purchase histories to identify natural groups of customers with similar buying patterns. They do not have predefined categories. Which type of machine learning should they use?

⚠ Common exam trap

A common mix-up: candidates confuse unsupervised clustering with supervised classification because both involve grouping, but classification requires predefined labels while clustering discovers groups from unlabeled data.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

Unsupervised clustering

Unsupervised clustering is the correct approach because the company wants to discover natural groupings in customer purchase histories without predefined labels. Clustering algorithms, such as K-Means or DBSCAN, partition data into clusters based on feature similarity, enabling the identification of customer segments with similar buying patterns without any prior training on labeled examples.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

  • Reinforcement learning

    Why it's wrong here

    Reinforcement learning trains an agent to choose actions through a trial-and-error interaction loop, using reward or penalty signals to maximize cumulative return over time. Customer purchase histories are a static, historical dataset with no live environment, no action-defining agent, and no reward feedback per transaction. While reinforcement learning could optimize future promotions, it cannot directly discover customer segments in existing unlabeled transaction data.

  • Supervised classification

    Why it's wrong here

    Supervised classification would require a pre-existing label on each customer record, such as 'high value' or 'segment A', to train a model that maps features to those known classes. The scenario only provides raw purchase histories with no such ground-truth labels or predefined customer categories. Because classification cannot invent new labels or discover groupings not present in the training data, it is unsuitable here.

  • Unsupervised clustering

    Why this is correct

    Unsupervised clustering is the appropriate technique because it discovers hidden structure in unlabeled data by grouping customers with similar purchase patterns. Algorithms such as k-means or DBSCAN can segment customers based on features like purchase frequency, total spend, and product-category preferences without any prior labels. This directly answers the business goal of analyzing purchase histories to identify natural customer segments.

  • Supervised regression

    Why it's wrong here

    Supervised regression models predict a continuous numeric outcome, such as the expected dollar amount of a customer's next purchase or their lifetime value. The business question here asks for categorical groupings, not a single numeric prediction, and regression also requires labeled target values for training. Even if labels existed, regression does not produce cluster assignments, so it cannot identify distinct customer segments.

About these practice questions

One of 985 original AI-900 practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →

How Courseiva writes practice questions · Editorial policy

JA

Written by Johnson Ajibi, MSc IT Security

Senior Network & Security Engineer · founder of Courseiva

This AI-900 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI-900 exam.