Types of Learning

From WikiOD

Supervised Learning[edit | edit source]

The machine learns to predict an output when given an input.

Each training case consists of an input and a target output.

Regression[edit | edit source]

The target ouput takes continuous values.

  • Predicting the price of a stock
  • Predicting a house price

Classification[edit | edit source]

The target output is a class label.

  • What kind of fruit the input is

Reinforcement Learning[edit | edit source]

The machine has to automatically determine the ideal behavior to maximize its performance.

For example:

Simple representation of a reinforcement learning algorithm

Using reinforcement learning you can also make a computer program that can complete a Mario level (MarI/O - Machine Learning for Video Games).

Unsupervised Learning[edit | edit source]

Unsupervised learning allows us to approach problems with little or no idea what our results should look like. We can derive structure from data where we don't necessarily know the effect of the variables.

The most common type of unsupervised learning is cluster analysis or clustering. It is the task of grouping a set of objects in such a way that object in the same group (cluster) are more similar to each other than to those in other groups.

There is also non-clustering unsupervised learning. An example thereof is identifying individual voices and music from a mesh of sounds. This is called the "Cocktail Party Algorithm".