K means (Cluster Analysis)
Contents
FCS Express can perform cluster analysis using k-means methodology.
Cluster analysis aims to group a set of objects/events in such a way that objects/events in the same group (i.e. a cluster) are more similar (in some sense or another) to each other than to those in other groups (i.e. the other clusters).
k-means is a partitioning-based clustering algorithm. k-means method for clustering is an iterative process in which an initial partition of given k clusters is then improved by applying a search algorithm to the data. Simplifying, given a pre-defined number (k) of clusters, the algorithm:
-begins with an initial set of k cluster centers (i.e. the centroids)
-(re)assigns objects to the closest centroids
-recalculates centroids according to new memberships of the data points.
-repeats the last two steps until a consistent result is found or until the maximum number of iterations is reached.
The basic k-means clustering is based on a non-deterministic algorithm. This means that running the algorithm several times on the same data, could give different results. However, to ensure consistent results, FCS Express performs k-means clustering using a deterministic method.
To begin working with cluster analysis in FCS Express please see the topics on:
Defining a k-means Clustering Analysis