You can learn a SOM for dimension reduction (operator name "SOMDimensionalityReduction") or train a SOM for Visualization (Plotview of ExampleSet). However, I cannot judge the quality of those. SOM for clustering is not available as far as I know.
Hi, isn't SOM Clustering equivalent to using SOM dimensionality reduction with target number of clusters as number of nodes and afterwards using kMeans on the reduced dimension? This would give you clusters containing all examples assigned on one node of the SOM grid.
Contributor I
Answers
You can learn a SOM for dimension reduction (operator name "SOMDimensionalityReduction") or train a SOM for Visualization (Plotview of ExampleSet). However, I cannot judge the quality of those. SOM for clustering is not available as far as I know.
regards,
Steffen
isn't SOM Clustering equivalent to using SOM dimensionality reduction with target number of clusters as number of nodes and afterwards using kMeans on the reduced dimension?
This would give you clusters containing all examples assigned on one node of the SOM grid.
Greetings,
Sebastian