Cluster analysis with SPSS Hierarchical Cluster Analysis：用SPSS聚类分析聚类分析.Image.Marked.pdf

Cluster analysis with SPSS: Hierarchical Cluster Analysis From the main menu consecutively click Analyze → Classify →Hierarchical Cluster. Figure 1. The following dialog window appears: Figure 2. Select the variables to be analyzed one by one and send them to the Variables box. Later actions greatly depend on which type of clustering is chosen here. For this purpose from the Cluster box we can choose between Cases, where we are performing clustering of the objects and Variables, where we are performing clustering of the variables. Next we must set the method for identifying the objects. The Label Cases by box is used for entering a string variable which labels the units. If instead of Cases (objects) we set Variables in the Cluster box, then we are required to set the variables in the Variable(s) list, and the Label Cases box is left empty. The default settings for the Display box are Statistics, for displaying the statistic results from the analysis and Plots, for displaying graphs. For both cases it is not necessary to remove the ticks. Four buttons for entering additional commands are located in the lower part of the dialog window. We activate the Statistics button (figure 2), which is used for defining the statistic results displayed on the screen. Here we can tick Agglomeration schedule box for displaying the agglomeration schedule or Proximity Matrix for displaying the proximity matrix, which presents the information for the distances between the objects and the clusters. In the first case, the sequence of unification of the objects in clusters is visualized and in the beginning each object is considered as a separate cluster and then the clustering is initiated. Further down in the Cluster Membership box we can choose: None - if it is not necessary to display the cluster membership of the objects; Single solution