AWS Certified Cloud Practitioner

Explanation

Unsupervised learning is the correct methodology for this scenario because:

1. Unlabeled Data: The company has "petabytes of unlabeled customer data." Unsupervised learning algorithms work with unlabeled data to find patterns, structures, or groupings within the data.

2. Customer Classification/Tiering: The goal is to "classify its customers into tiers." This is a classic clustering problem where unsupervised learning algorithms like K-means, hierarchical clustering, or DBSCAN can group similar customers together based on their characteristics without predefined labels.

3. Why Not Other Options:

   • A. Supervised learning: Requires labeled data (pre-classified examples) to train models. Since the data is unlabeled, supervised learning isn't appropriate.
   • C. Reinforcement learning: Focuses on learning optimal actions through trial-and-error interactions with an environment to maximize rewards. Not suitable for customer classification from static data.
   • D. Reinforcement learning from human feedback (RLHF): A specialized form of reinforcement learning that incorporates human feedback, typically used for fine-tuning language models or complex decision-making systems. Not applicable for customer segmentation from unlabeled data.

Key Takeaway: When you have unlabeled data and need to discover patterns or groupings within it, unsupervised learning is the appropriate machine learning approach.