Module:  K-means Clustering                                                                                       Print

Lesson Objective Learn how to divide rows of data into separate groups based upon multivariate pattern similarities
Section Description
K-means Clustering A basic discussion of K-means clustering and the notion of the underlying similarity calculations.  How to launch the K-means clustering tool.
K-means Clustering Dialog Box Determining the columns and rows to be used for the calculations and methods for dealing with empty values.  Input options for the maximum number of clusters and similarity measure.
Cluster Initialization Methods A detailed discussion of each of the available cluster initialization methods, including:  data centroid based search, evenly spaced profiles, randomly generated profiles, randomly selected profiles, and from marked records.
K-means Clustering Results Examining the new columns, new visualizations, and K-means stats in the Legend window.
Evaluating K-means Clustering Results A recommended approach to evaluating the quality of your clustering by examining the similarity to centroids and diversity between cluster groupings.