Why Self-Organizing Map Neural Networks?
Self-organizing maps (SOMs) are very good at creating classifications. Further, the classifications retain topological information about which classes are most similar to others. Self-organizing maps can be created we any desired level of detail. They are particularly well suited for clustering data in many dimensions and with complexly shaped and connected feature spaces. They are well suited to cluster iris flowers. The four flower attributes will act as inputs to the SOM, which will map them onto a 2-dimensional layer of neurons.
