Definition:Cluster Analysis

From ProofWiki
Jump to navigation Jump to search

Definition

Cluster analysis is a branch of statistics which determines whether, on the basis of observations, the data falls into clusters.

For a given metric, items in one cluster will in general be closer to each other than to items in another cluster.

Cluster analysis provides objective criteria to determine the number of clusters and to determine which items go into which cluster.


Examples

Human Statistics

When analysing a human population, various statistics such as:

  • age
  • income
  • ownership of home or car
  • time spent out of the house each week

are likely to show evidence for several distinct clusters, each corresponding to a category such as:

  • unemployed
  • pensioners
  • students

and so on.


Also see

  • Results about cluster analysis can be found here.


Sources