Non-hierarchical Cluster Analysis Options

This menu controls aspects of the way in which the clustering is carried out, and selects the information to be printed.

Display

Specifies what is printed during the cluster analysis.
Criterion valueoptimal criterion value
Classified datadata with the units ordered into the optimal classes
Typical value for each classdisplays a typical value for each class: for maximal predictive classification this is the class predictor; for the other methods it is the class mean.
Optimum classificationthe optimum classification
If Display Values from Initial Classification is checked, the requested sections of output are also displayed for the initial classification.

Initial Classification

This allows you to select how the initial classification is formed. Equal-sized groups by unit order splits the units, in order, into k groups of nearly equal size. Automatic, by Distance, finds the k units that are furthest apart in the multi-dimensional space defined by the data variates; these are then used as nuclei for the classes, with each remaining unit being allocated to the class containing the nearest nucleus.

Between-Group Interchanges

This specifies which types of interchange are to be used to optimize the classification: both transfers and swops, only swops, or whether the clusters are to be fixed at those given by the initial classification.