Clustering - Definition and Overview

Clustering can refer to

• (in Computer science) the connection of many low-cost computers using special hardware and software such that they can be used as one larger computer.

The hardware used to link nodes together is usually called interconnect and consists of Network Interface Cards, or NICs, Switches and cables. The choice of interconnect depends on many factors. For fine-grained applications, where the nodes need to communicate often, the interconnect must be fast or the application suffers waiting for the network.

In terms of Software, clustering can either be used to provide reliability (when one machines fails, the other(s) take over its workload) or as a means to inexpensively provide large amounts of computing power. See computer cluster.

• a common technique for data analysis, which is used in many fields, including machine learning, data mining, pattern recognition, image analysis,bioinformatics, and search engines. Clustering consists of partitioning a data set into subsets (clusters), so that the data in each subset (ideally) share some common trait - often similarity or proximity for some defined distance measure. When the data is human language text, rather than numbers or symbols, different methods are employed. See Data clustering
• (In demographics) to the gathering of various populations based on ethnicity, economics or religion. See Clustering (demographics).