Categories

Versions

Extract Cluster Prototypes (AI Studio Core)

Synopsis

This operator generates an ExampleSet consisting of the Cluster Prototypes from the Cluster Model. This operator is usually applied after clustering operators to store the Cluster Prototypes in form of an ExampleSet.

Description

Most clustering algorithms like K-Means or K-Medoids cluster the data around some prototypical data vectors. For example the K-Means algorithm uses the centroid of all examples of a cluster. The Extract Cluster Prototypes operator extracts these prototypes and stores them in an ExampleSet for further use. This operator expects a cluster model as input. The information about the cluster prototypes can be seen in the cluster models generated by most clustering operators but the Extract Cluster Prototypes operator stores this information in form of an ExampleSet thus it can be used easily.

Clustering is concerned with grouping together objects that are similar to each other and dissimilar to the objects belonging to other clusters. Clustering is a technique for extracting information from unlabeled data. Clustering can be very useful in many different scenarios e.g. in a marketing application we may be interested in finding clusters of customers with similar buying behavior.

Differentiation

k-Medoids

The K-Medoids operator performs the clustering and generates a cluster model and a clustered ExampleSet. The cluster model generated by the K-Medoids operator can be used by the Extract Cluster Prototypes operator to store the Centroid Table in form of an ExampleSet.

k-Means

The K-Means operator performs the clustering and generates a cluster model and a clustered ExampleSet. The cluster model generated by the K-Means operator can be used by the Extract Cluster Prototypes operator to store the Centroid Table in form of an ExampleSet.

Input

  • model (Centroid Cluster Model)

    This port expects a cluster model. It has information regarding the clustering performed by a clustering operator. It tells which examples are part of which cluster. It also has information regarding centroids of each cluster.

Output

  • example set (Data table)

    The ExampleSet consisting of the Cluster Prototypes is generated from the input Cluster Model and the ExampleSet is delivered through this port

  • model (Centroid Cluster Model)

    The cluster model that was given as input is passed without any changes to the output through this port.

Tutorial Processes

Extracting Centroid Table after application of the K-Means operator on Ripley-Set

In many cases, no target attribute (i.e. label) can be defined and the data should be automatically grouped. This procedure is called Clustering. There is a wide range of clustering schemes supported which can be used in just the same way like any other learning scheme. This includes the combination with all preprocessing operators.

In this Example Process, the 'Ripley-Set' data set is loaded using the Retrieve operator. Note that the label is loaded too, but it is only used for visualization and comparison and not for building the clusters itself. A breakpoint is inserted at this step so that you can have a look at the ExampleSet before application of the K-Means operator. Other than the label attribute the 'Ripley-Set' has two real attributes; 'att1' and 'att2'. The K-Means operator is applied on this data set with default values for all parameters. Run the process and you will see that two new attributes are created by the K-Means operator. The id attribute is created to distinguish examples clearly. The cluster attribute is created to show which cluster the examples belong to. As parameter k was set to 2, only two clusters are possible. That is why each example is assigned to either 'cluster_0' or 'cluster_1'. A cluster model is delivered through the cluster model output port. It has information regarding the clustering performed. Under Folder View you can see members of each cluster in folder format. You can see information regarding centroids under the Centroid Table and Centroid Plot View tabs. A breakpoint is inserted at this step so that you can have a look at the cluster model (especially the Centroid Table) before application of the Extract Cluster Prototypes operator. The Extract Cluster Prototypes operator is applied on the cluster model generated by the K-Means operator which stores the Centroid Table in form of an ExampleSet which can be seen in the Results Workspace.