Lesson 5 of 7 · 12 min
Supervised, unsupervised and deep learning
Supervised learning learns from labelled examples to predict, unsupervised learning finds structure in unlabelled data, and deep learning uses many-layered neural networks, with either approach, to recognise complex patterns.
In short
- Supervised learning: inputs and outputs are labelled; the model learns the mapping and then predicts outcomes for new data.
- Unsupervised learning: no labels; the algorithm describes the data and their structure, such as grouping firms into peer groups.
- Deep learning: neural networks with many hidden layers doing multistage, non-linear processing; it can be supervised or unsupervised.
- Deep learning builds from simple concepts to complex ones and excels at image, pattern and speech recognition.
- Investment uses include forecasting market direction, choosing signals, predicting mergers or elections, and reading satellite images of car parks, ports, factories and crops.
Unlock this lesson free for 7 days
Create a free account to get 7 days of full access — every lesson, video, flashcard, mock and the question bank. No card needed.