usecase: Clustering Kohonen Network

public usecase: Clustering Kohonen Network
Project: Phase: 1.0; Status: Proposed; Version: 1.0; Complexity: 1
Dates: Created: 9.12.2005 16:46:32; Modified: 14.12.2005 8:25:24;
Flags: Active: false; IsRoot: false; IsLeaf: false;
Extension Points:
UUID: {C3B3ABA1-14F8-4120-9B3D-6373B732FC3E}

Goto: Scenarios

See also: User

Appears in: UC Clustering Network

Clustering Kohonen Network Scenarios
Scenario Type Detail
Clustering Basic Path First is generating array of KNeuron, that need clustering algorithms. After that are retrieved settings for clustering from SOMForm.clusterSettings class.

For hierarchical clustering is used: DataClustering.HierarchicalClustering(IClusterable[] data, int endClusters, int distance) function
For K-Mean clustering is used:
DataClustering.KMeanClustering(IClusterable[] data, int endClusters) function

After clustering, every neuron in network has clusterID ( KNeuron.getClusterID() ), that derermine to which cluster neuron belongs to.