An email must be labeled as spam or not spam. Which technique fits?
Why D is the answer
Classification
Classification assigns one of a set of labels.
In detail
- Spam or not spam is a label, so this is classification.
- Two possible labels is called binary classification.
Use case
An illustrative scenario, not a real incident.
A mail provider labels each message as spam or not spam, which is classification.
See it in different ways
flowchart TD
A{"What does the data look like?"} -->|"Labeled examples"| B["Supervised: classification, regression"]
A -->|"No labels"| C["Unsupervised: clustering"]
A -->|"Rewards from actions"| D["Reinforcement learning"]mindmap
root((Machine learning))
Supervised
Regression: a number
Classification: a label
Unsupervised
Clustering
Reinforcement
Learns from rewardsflowchart LR S0["Email"] --> S1["Classifier"] --> S2["Spam or not spam"]
Key point: Classification assigns labels.
Why the other options are wrong
- A. Clustering
- Clustering has no labels.
- B. Topic modeling with no labels
- Topic modeling has no labels.
- C. Regression
- Regression predicts a number.