Types Of Machine Learning
Supervised, Unsupervised & Semi-Supervised Learning
Short description explaining what the learner will understand after completing this study map.
Three Ways to Learn from Data
Machine learning approaches can be distinguished by the information available during learning. Supervised learning uses labelled examples, unsupervised learning works with unlabelled data, and semi-supervised learning combines a smaller labelled set with a larger unlabelled set.
Core idea: look first at the training data. Are correct outputs supplied, absent, or available for only part of the dataset? That distinction points toward supervised, unsupervised, or semi-supervised learning.
Essential Learning Terms
Tap each card to review the concepts that distinguish the three learning approaches.
Supervised Learning
Supervised learning trains on labelled examples. Each training example includes input information together with a known target or correct output, allowing the model to learn a relationship between inputs and outputs.
Unsupervised vs Semi-Supervised
Unsupervised learning receives data without target labels and searches for useful structure or patterns. Semi-supervised learning sits between the two approaches by using both labelled and unlabelled examples.
Unsupervised Learning
Works with unlabelled data. A common goal is to discover structure, such as groups of similar observations through clustering.
Semi-Supervised Learning
Uses a mixture of labelled and unlabelled data, which can be useful when obtaining labels is more difficult or costly than collecting raw examples.
Compare the Learning Approaches
The presence of labels is the fastest way to distinguish these approaches. The learning objective then helps refine the choice: predict known outputs, discover structure, or learn from a mixture of labelled and unlabelled examples.
UNSUPERVISED → UNLABELLED DATA
SEMI-SUPERVISED → LABELLED + UNLABELLED DATA
Recognizing the Approach
A house-price model trained on examples containing known sale prices is supervised. Customer segmentation performed without predefined customer categories is unsupervised. A classifier trained with a small labelled collection plus many unlabelled examples is semi-supervised.
Learning Approach Explorer
Explore how labels and training data determine the learning approach.
Key Takeaways
Focus on the training information: all examples labelled → supervised; no target labels → unsupervised; some labelled and many unlabelled → semi-supervised.
Core pattern: supervised learning learns from labelled examples; unsupervised learning discovers structure in unlabelled data; semi-supervised learning combines labelled and unlabelled examples.
Common mistake: supervised does not mean a person watches every prediction. It means the learning process uses training examples with known target values or labels.
Knowledge Check
These questions test whether you can distinguish supervised, unsupervised, and semi-supervised learning from the data and task described.