Unsupervised Learning

Unsupervised learning is a machine learning technique where the machine learns from experience without having reference examples and answers.

Note. In supervised learning, the machine learns from examples and solutions provided by a supervisor. In reinforcement learning (RL), on the other hand, the machine learns through a reward function (reinforcement).

In unsupervised learning, the data is unlabeled.

The structure of the data itself is not predefined.

To learn, the machine must extract relevant information from the available data.

Techniques of Unsupervised Learning

The main techniques of unsupervised machine learning include:

• Clustering. The learning algorithm looks for patterns in the available data. It's particularly useful in big data analysis.
  
• Data Dimensionality Reduction. The learning algorithm eliminates insignificant data (noise) and combines redundant information (correlations) to focus the analysis on data that reveals a pattern.