The difference is whether the answers are given.
Direct answer
In supervised learning you train a model on labelled examples, such as emails marked spam or not spam, so it predicts labels for new data. In unsupervised learning there are no labels and the model finds structure on its own, such as grouping similar customers. Reinforcement learning learns from rewards.
Comparison
| Supervised | Unsupervised | |
|---|---|---|
| Data | labelled | unlabelled |
| Goal | predict | discover structure |
| Examples | spam detection, price prediction | customer segments, anomaly detection |
| Algorithms | regression, decision trees | clustering, dimensionality reduction |
Two kinds of supervised learning
- Classification: predict a category.
- Regression: predict a number.
How to choose
- Do you have labels? Try supervised.
- Exploring unknown data? Try unsupervised.
- Learning by trial and feedback? Consider reinforcement learning.
Where Teyro fits
Teyro's Machine Learning for Beginners course turns this into short lessons you can play in a few minutes a day. It has 200 lessons, and you can start free.
The bottom line
Labels or no labels decide the approach.
