Topic Status: Some information in this topic is preview and subject to change in future releases. Preview information describes new features or changes to existing features in Microsoft SQL Server 2016 Community Technology Preview 2 (CTP2).
Use the Cluster Discrimination tab to compare two clusters that exist in a clustering model. You can see how different combinations of attributes and values are represented within the clusters.
Choose a mining model from those in the current mining structure. The mining model will open in its associated viewer.
Choose a viewer to use to explore the selected mining model. You can use the custom viewer for clustering models, or the Microsoft Mining Content Viewer. You can also use plug-in viewers if available.
Select a cluster, so that you can compare it to another cluster.
Select a second cluster from the list of clusters in the mining model, to compare to Cluster 1. You can also compare a cluster to its complement, meaning all cases in the model except those in the selected cluster.
Discrimination scores for <cluster 1> and <cluster 2>
The columns in the graph provide information about how each attribute-value pair is related to the two selected clusters.
An attribute in the mining model.
A value of the attribute selected in Variables.
Favors <cluster 1>
The bar graph on the left represents the probability that the selected attribute-value pair is representative of the cluster selected in Cluster 1. You can pause the mouse over the bar to see the value, represented as a percentage. Note that even if the value is zero, it doesn’t mean the attribute-value is necessarily missing from the cluster, just that the distribution strongly favors one cluster over the other.
Favors <cluster 2>
The bar graph on the right represents the probability that the selected attribute-value pair is representative of the cluster selected in Cluster 2.