AWS Certified AI PractitionerAI/ML and Generative AI FundamentalsEasy
A financial institution is implementing an AI system to detect anomalies in transaction data. The system needs to identify unusual patterns without prior examples of fraudulent activity, as new types of fraud constantly emerge. Which type of machine learning approach is best suited for this scenario?
- AUnsupervised Learning
- BSupervised Learning
- CSemi-supervised Learning
- DReinforcement Learning
Show answer & explanationAnswer & explanation
Correct answer: A. Unsupervised Learning
Unsupervised learning is ideal for anomaly detection when there are no pre-labeled examples of anomalies. Algorithms like clustering or autoencoders can learn the 'normal' patterns in data and then flag data points that deviate significantly from these learned patterns as anomalies, which is perfect for detecting new, unknown types of fraud.
Why the other options are wrong
- B. Supervised learning requires labeled data (examples of both normal and fraudulent transactions), which is not available for new types of fraud.
- C. Semi-supervised learning uses a small amount of labeled data with a large amount of unlabeled data, but the scenario explicitly states 'without prior examples of fraudulent activity'.
- D. Reinforcement learning is for learning optimal actions in an environment and is not typically used for anomaly detection.
Unsupervised Learning for Anomaly Detection
Unsupervised learning is an ideal approach for anomaly detection when there are no pre-labeled examples of abnormal data. It works by learning the normal patterns in the data and flagging observations that deviate significantly from these patterns.
- Does not require labeled anomaly data.
- Identifies outliers or novel patterns.
- Common algorithms include Clustering (e.g., K-Means, DBSCAN) and Autoencoders.
Memory trick: Finding the weird without knowing what 'weird' looks like.