Clustering Models
This article is part of a series.
http://social.technet.microsoft.com/wiki/contentfiji-r1">
This article is part of a series.
http://social.technet.microsoft.com/wiki/contentfiji-r1">
Returns the characteristics of a specific node that represents a cluster.
System.Microsoft.AnalysisServices.System.DataMining.Clustering.GetClusterCharacteristics
Model name as string
Count of attributes as integer
ID of attribute to start at as integer
This example gets the characteristics of the cluster with the ID of 9, where the frequency score for each characteristic of the cluster is greater than 0.5.
CALL System.GetClusterCharacteristics('TM Clustering', '009',0.5)
Attributes |
Values |
Frequency |
Support |
Region |
North America |
0.973831946 |
1047 |
House Owner Flag |
0 |
0.890332939
|
957 |
Marital Status |
S |
0.837178021 |
957 |
Marital Status |
900 |
||
Bike Buyer |
1 |
0.567955122 |
610 |
This stored procedure is used by the Microsoft Clustering Viewer.
The following article provides additional examples of how to work with clustering models.
http://www.sqlserverdatamining.com/ssdm/Default.aspx?tabid=102&Id=32
Compares a cluster with another cluster or with the complement of the cluster.
System.Microsoft.AnalysisServices.System.DataMining.Clustering.GetClusterDiscrimination
Model name as string
Unique node ID as string
Second unique node ID as strig
Probability threshold value as double
Rescaled as Boolean
This example compares cluster 9 with its complement, meaning all other cases in the model. The value 0.01 in the arguments filters the results, so that only attributes with a probability greater than that score are returned.
CALL System.GetClusterDiscrimination('TM_Clustering','009','',0.01,true)
In this example, a positive score means that the attribute favors the specified cluster; a negative score means that the attribute favors the complement.
Attributes |
Values |
Scores |
Region |
North America |
-1846.71 |
Region |
Europe |
1130.736 |
Occupation |
Manual |
1063.709 |
Occupation |
Skilled Manual |
-965.511 |
This example compares cluster 9 to cluster 8. A positive score in the results means that the attribute favors the cluster 9; a negative score means that the attribute favors cluster 8.
The rescaled parameter false indicates that probabilities should not be normalized.
CALL System.GetClusterDiscrimination('TM Clustering', '009','8',0.2,false)
If a second node is not specified for comparison, the procedure returns the complement of the first node, meaning all cases not in the first specified node.
If you specify a node ID of 0, the attributes for the entire model are returned.
The following article provides additional examples of how to work with clustering models.
http://www.sqlserverdatamining.com/ssdm/Default.aspx?tabid=102&Id=32
Returns the characteristics of the cluster, with all the information that is found in the NODE_DISTRIBUTION table for the cluster.
System.Microsoft.AnalysisServices.System.DataMining.Clustering.GetClusterProfiles
Model name as string
Model name as String
Count of attributes as integer
ID of attribute to start at as integer
This example gets the profiles of the clusters for the two attributes beginning at attribute ID 1, because attribute IDs are 0-based.
CALL System.GetClusterProfiles('TM Clustering',2,1)
This stored procedure executes a series of DMX statements that return the complete set of profiles in the model. You must know the IDs of the attributes that you want to include in the results.
The following article provides additional examples of how to work with clustering models.
http://www.sqlserverdatamining.com/ssdm/Default.aspx?tabid=102&Id=32
The following article describes the content of the NODE_DISTRIBUTION node in more detail.
Layout of Clustering Algorithm Content
The following article provides additional examples of how to work with clustering models.
http://www.sqlserverdatamining.com/ssdm/Default.aspx?tabid=102&Id=32
m>
Returns a list of all clusters in the model, with the node ID and the name of each cluster.
System.Microsoft.AnalysisServices.System.DataMining.Clustering.GetClusters
Model name as string
This example returns a list of all clusters in the model, with IDs.
CALL GetClusters('TM Clustering')
ClusterName |
ClusterUniqueNodeID |
Top customers |
001 |
Cluster 2 |
002 |
Cluster 3 |
003 |
If you’ve renamed any of the clusters, the new names are returned.
The following article provides additional examples of how to work with clustering models.
http://www.sqlserverdatamining.com/ssdm/Default.aspx?tabid=102&Id=32