Analyze > Classify > Hierarchical Cluster opens a dialogue that lets you specify the variables you want to analyse. When meaningful select a variable to label observations on the plots produced. The procedure can be used to cluster either variables or observations (cases).

Method This button lets you choose the clustering method to be used, as well as to specify the distance measures.

The following cluster methods are available:

Many distance measures are provided for various types of variables

Save offers options to copy group membership variables to the current data set. You can either specify a single solution (how many clusters the selected solution should have) or a range e.g 2 to 5 clusters.

Advice: Examine the dendgrogramme to choose the number of clusters.

Plots Specifies the plots you wish to produce: dendrogrammes or icicle plots (with a number of options).

Please refer to the SPSS documentation for details (CLUSTER command).

• K-Means Clustering [Quick Cluster]
• Two step cluster
• Decision trees