Generalized Cluster Analysis Results - Quick tab

Select the Quick tab of the Generalized Cluster Analysis results dialog box to access the options described here. For more details regarding the various results options, see also the Introductory Overview.

Element Name Description
Cluster means This option only applies to k-Means clustering (see also the Introductory Overview for details). Click the Cluster means button to produce a spreadsheet that contains the cluster means and the number of cases and percentage of cases in each cluster.
Cluster distances This option only applies to k-Means clustering. Click the Cluster distances button to produce a spreadsheet with distances between all cluster centroids. See also the Generalized Cluster Analysis - k-Means tab for a description of the distance measures available for k-Means clustering.
Graph of continuous variable means Click this button to produce a line graph that shows the scaled cluster means for all continuous variables. These means are computed as follows:

where

is the transformed (scaled) mean for continuous variable i and cluster j

 is the arithmetic ("unscaled") mean for continuous variable i and cluster j

are the maximum and minimum observed values for continuous variable i

In other words, the plotted values depict the means scaled to the overall ranges of observed values for the respective continuous variables.

Graph of cost sequence Click this button to produce a line graph showing the respective error function (average distance of cases to cluster centers for k-Means; average log-likelihood of cases for EM) over the range of cluster solutions considered by the program. Note that this graph is only available when v-fold crossvalidation is used. See also the Introductory Overview for more details.
Cluster properties Use the options in this group box to review the assignment of observations (cases) to clusters, and other statistics computed for observations (distances from centroids for k-Means clustering; classification probabilities for EM clustering).
Drop-down list Select the specific cluster for which to review the respective results; leave the default All clusters to review the results for all cases.
Sort by cluster number Select this check box to display the results for All clusters sorted by (assigned) cluster, rather than case number.
Members & distances This option applies only to k-Means clustering. Click the Members & distances button to produce a spreadsheet containing the final cluster assignments for all cases, the input data (variable values), and the distances of the cases to the respective (assigned) cluster centers. If a Sample Identifier Variable is specified in the Cross-Validation dialog box, you can select to display these values for the Analysis sample only, the Test set sample, or both (All samples), depending on which option is selected in the Sample group box.
Save classifications & distances This option only applies to k-Means clustering. Click this button to display a standard variable selection dialog box, which is used to select variable(s) to be displayed together with the final cluster assignments (classifications) and distances to the respective cluster centroids. After you select variable(s), a spreadsheet containing these values will be displayed in an individual window (regardless of the settings on the Options dialog box - Output Manager tab or the Analysis/Graph Output Manager dialog box).
Priors (a-priori probabilities) This option applies only to the EM method of clustering. Click the Priors (a-priori probabilities) button to display a spreadsheet that contains the a priori probability values for each cluster.
Classification probabilities (weights) This option applies only to EM clustering. Click the Classification probabilities (weights) button to display a spreadsheet that contains the classification probabilities for each observation and cluster, along with the final classifications (cluster assignment to the cluster with the highest probability).
Members & probabilities This option applies only to EM clustering. Click the Members & probabilities button to display a spreadsheet containing the final cluster assignments for all cases, the input data (variable values), and the classification (cluster assignment) probabilities of the cases for the respective (assigned) clusters. If a Sample Identifier Variable is specified in the Cross-Validation dialog box, you can select to display these values for the Analysis sample only, the Test set sample, or both (All samples), depending on which option is chosen in the Sample group box.
Save classifications & probabilities This option applies only to EM clustering. Click this button to display a standard variable selection dialog box, which is used to select variable(s) to be displayed together with the final cluster assignments (classifications) and probabilities. After you select these variable(s), a spreadsheet containing these values will be displayed in an individual window (regardless of the settings on the Options dialog box - Output Manager tab or the Analysis/Graph Output Manager dialog box).