Unsupervised learning
Unsupervised learning is a type of machine learning algorithm which does not supervise the model, but lets the model work on its own. The biggest difference between supervised learning and unsupervised learning is that the former deals with labeled data, while the latter deals with unlabeled data. While supervised learning is about function approximation, unsupervised learning is about description. Unsupervised learning has more difficult algorithms than supervised learning. Its goal is to automatically find structure in a dataset, to find the regularities in the input, to see what normally happens. Unsupervised learning assumes there is a structure to the input space which implies there are certain patterns that occur more often than others.
Applications
The main applications of unsupervised learning are listed as follows.[1]
- Clustering: The process of grouping the given data into different clusters or groups.
- Visualization: The process of creating diagrams, images, graphs, charts, etc., to communicate some information.
- Dimensionality Reduction: The process of reducing the number of random variables under consideration by getting a set of principal variables.
- Finding Association Rules: The process of finding associations between different parameters in the available data.
- Anomaly Detection: The identification of rare items, events or observations which brings suspicions by differing significantly from the normal data.