Kohonen Network (SOM)

The Kohonen self-organizing map (SOM) network performs a mapping from input vectors to a output vectors, preserving the topological properties of the input. This means that vectors close to each other in the input space are mapped to the same or neighboring neurons in the output space. It can be used for visualisation or sorting of high dimensional data. [http://en.wikipedia.org/wiki/Kohonen]

To create and train Kohonen neural network with easyNeurons do the following:

  1. Choose Kohonen architecture (in main menu choose Networks>Kohonen)
  2. Enter architecture specific parameters (number of neurons in input layer)
  3. Create training set (in main menu choose Training >New Training Set)
  4. Train network
  5. Test network

Step 1. To create Kohonen network, in main menu click Networks > Kohonen

 Step 2. Enter number of neurons in input layer, and click Create button.

This will create the Kohonen neural network with two neurons in input and nine in output layer.

Now we shall train this simple network, to learn from data. First we have to create the training set

Step 3.  In main menu click Training > New Training Set to open training set wizard.

Step 4. Train network

TODO

Step 5. Test network

TODO

 

See also

http://www.learnartificialneuralnetworks.com/kohonen.html