AIF-C01 ยท AI use cases and AWS AI services
Regression, classification and clustering
Matching a business question to regression, classification or clustering, and real-world applications such as forecasting, fraud detection and recommendations.
Key points
- Regression predicts a number (price, demand forecast).
- Classification predicts a category from labeled examples (fraud or not fraud, spam or not spam).
- Clustering groups similar items without labels to discover structure (customer segments).
- Ask what the output looks like: a number, a label, or an unlabeled grouping.
Practice questions
Which technique predicts a continuous numeric value such as a price?
- Regression
- Classification
- Clustering
- Tokenization
Show answer
Answer: A. Regression
Regression outputs a number on a continuous scale. Classification outputs a category and clustering outputs groups without labels.
Which tasks are classification problems?
- Marking an email as spam or not spam
- Flagging a card transaction as fraudulent or legitimate
- Estimating next month's revenue in euros
- Discovering customer segments without any labels
- Predicting delivery time in minutes
Show answer
Answer: A. Marking an email as spam or not spam, B. Flagging a card transaction as fraudulent or legitimate
Spam detection and fraud flagging assign each case to a category. Sales forecasting and delivery time are numeric (regression); unlabeled segmentation is clustering.
True or false: clustering requires a labeled example for every group it should find.
Show answer
Answer: False
False. Clustering is unsupervised: it finds groups from similarity alone. Needing labeled examples is a trait of classification.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
