The Cluster Diagram tab in the Microsoft Sequence Clustering Viewer provides a graphical view of all the clusters that the sequence clustering model contains.
Use this view of a sequence clustering model to drill through from each cluster to the supporting cases, if drillthrough has been enabled. You can also assign descriptive names to the clusters, and change the shading variable to assess the distribution of values at a glance
Note that in this context, you are searching only the name of the cluster, not the attributes within the cluster; therefore, this option is most useful if you have assigned descriptive names to your cluster. You can assign names to clusters in the viewer by right-clicking each cluster and selecting Rename.
Reorder the clusters in the diagram to improve the layout.
The appearance of the density bar graph and the values in it are dependent on the attribute that you select in Shading Variable.
If no attribute state is selected as the shading variable, by default the density shading applied to each cluster represents the support for the cluster, compared to the overall population of cases.
When you select an attribute for Shading Variable, you must then select a value for State as well. When you do so, the density bar graph updates to show the probability of this state. You can pause the mouse over any individual cluster to see the probability of the selected state for the cluster.
Select an attribute from the mining model to use in shading the cluster diagram.
Select a state that corresponds to the Shading Variable. For example, if you want to view the sequences that include a particular product, you would select the [Product] column as the attribute for Shading Variable, and the specific product name as the State value.
The lines in the diagram indicate associations between sequence clusters. You can adjust how many links the viewer shows by adjusting the slider to the right of the clusters. Lowering the slider shows only the strongest links.