supervised training of SOM in MATLAB

Question

I have a labelled data set, each data entry is of six dimensional. Each data entry is pre-labelled as belonging to one of 10 clusters.

I would like to train a SOM to fit this labelled data set. In other words, I would like to enforce a SOM that can exactly( or almost) cluster the same result to each data entry as the pre-labelled one.

Is there a function in the MATLAB neural network toolbox that can fulfill the above requirement?

Answer 1

Unless I've misunderstood your question, the activity you're describing sounds to me more like classification (i.e. building a model in a supervised way, that will predict the class of new observations) than clustering (i.e. building a model in an unsupervised way to explore the dataset structure).

If that's right, then yes, Neural Network Toolbox has the functionality to do that, but it isn't SOM-based (SOMs are typically unsupervised).

nprtool (the neural network pattern recognition tool) is a GUI tool that comes with the toolbox for this activity. It has a very friendly wizard structure that walks you through the activity, and you can even select built-in example datasets to work with. It should make it clear how to carry it out. Once you've walked through an example in the GUI, you can generate code from the GUI which will show you how to do the equivalent thing at the command line.

Hope that helps!