som

Implementation of a self-organizing map.

Create a new map object with

tmpSom =  SOM.new(learningRate,maxIterations,xSize,ySize,weightCount)

Train the map with

tmpSom.train(dataSet)

where dataSet is a 2D-array of weightCount length vertices.

After training render a csv file with

tmpSom.export2cvs('filename')

make filename out to be whatever you please.

TODO:

• Speed up training. Possibly through multi-threading.
• Enable hexagonal neighbourhoods while training.
• Enable hexagonal shaped maps.
• Add comments to code
• Convert to use mongoDB to gain advantage of the shape selection query.