There are no user-supported procedures for generating these types of analyses in SAS directly, but SAS Enterprise Miner has an SOM/Kohonen Node for building Self-Organizing Maps or Kohonen Networks. They can be used to create segments much like clustering methods and it has been argued that these methods might be more stable than k-means clustering in many situations. It is also possible to be misled by the visual display created by such an analysis as it might direct your thinking along those lines. You might find it useful to analyze the data in several ways including using SOMs and clustering methods to better understand how the solutions differ with respect to a particular set of input data. You can find additional documentation for these nodes in the software itself by clicking on Help --> Contents and then navigating in the panel on the left to
Node Reference
Explore
SOM/Kohonen Node
and then clicking on Overview of the SOM/Kohonen Node in the panel on the right. From this section, you will find the following comments:
You use the SOM/Kohonen node to perform unsupervised learning by using Kohonen vector quantization (VQ), Kohonen self-organizing maps (SOMs), or batch SOMs with Nadaraya-Watson or local-linear smoothing. Kohonen VQ is a clustering method, whereas SOMs are primarily dimension-reduction methods. For cluster analysis, the Clustering node is recommended instead of Kohonen VQ or SOMs.
You can run SAS Enterprise Miner in batch mode but you cannot generate SOMs in SAS without having a SAS Enterprise Miner license.
Let me know what you think!
Doug
