AWS Certified Cloud Practitioner

Explanation

Unsupervised learning is the correct answer because:

1. No labels or prior classifications exist - This is the key characteristic of unsupervised learning problems. The algorithm must discover patterns and structures in the data without any predefined labels or target variables.

2. Customer segmentation - This is a classic unsupervised learning task where the goal is to group similar customers together based on their features (buying behavior in this case).

3. Comparison with other options:

   • Supervised learning (A): Requires labeled data with known outcomes to train models. Not suitable when no labels exist.

   • Reinforcement learning (C): Involves an agent learning through trial-and-error interactions with an environment to maximize rewards. Not appropriate for customer segmentation.

   • Self-supervised learning (D): A subset of unsupervised learning that creates its own supervisory signals from the data, but typically used for different types of problems like representation learning.

4. Common unsupervised techniques for segmentation:

   • Clustering algorithms (K-means, hierarchical clustering, DBSCAN)

   • Dimensionality reduction (PCA, t-SNE) for visualization

   • Association rule learning for discovering relationships between variables

Why this matters for AWS Cloud Practitioner: Understanding different machine learning approaches helps in selecting appropriate AWS services like Amazon SageMaker for ML tasks, and recognizing when to use unsupervised learning techniques for business problems like customer segmentation, anomaly detection, or market basket analysis.