Kohonen networks are a special type of artificial neural networks. These networks are able to learn data topology creating special maps called Self Organizing Maps (SOM). These maps are very useful during the clustering process because of their topological similarity to the training data.

Self Organizing Maps created by Kohonen Networks can be displayed. This means it is possible to obtain some kind of visualization even for very complicated data.

It is possible to use Kohonen Networks as a clustering method (the classical SOM algorithm), as well as a classification method (the LVQ family of algorithms). In both cases a Self Organizing Map is obtained, which can be visualized.