Document Type : Research Paper
Author
Member of Faculty, Islamic Azad University of Najaf Abad
Abstract
Cluster analysis is a multivariate statistical method that can be used to group a set of observations or variables.
The simplest method of cluster analysis is the K-Means non-hierarchical method used for observation clustering in which the number of clusters is unclear. (Johnson and D.Wichern 1988).
In this method, observations are put in the same clusters based on the least difference among the means. Thus, at first a number of different groups are formed, and each of the observations is consecutively put in different groups and the difference between its mean and the mean of the group is calculated, until it is eventually placed in the group with the least calculated difference (Milligan, G.W.1980). To measure the difference between the means of observations, the Euclidean distance method is used.