The Kohonen Self Organizing Map network is a type of unsupervised network, which has the ability to learn without being shown correct outputs in sample patterns. These networks are able to separate data into a specified number of categories.
The Kohonen Self Organizing Map network contains only two layers: an input layer and an output layer which has one neuron for each possible category. You can specify the maximum number of output categories by setting the number of neurons in the output layer in the Kohonen Architecture module. In other words, if you want your patterns to be categorized into at most N clusters, then the output layer will contain N neurons. The Kohonen network may decide your patterns have less than N categories, but N will be the maximum it can find.