(Please click "Reset" after the applet has loaded)
This applet demonstrates a Kohonen Feature Map neural net in three-dimensional space.
The task of this neural net is to span its map over all blue points (the input values of the net) in an even way and without crossing one point twice. This problem is similar to the classic Travelling Salesman Problem, where the shortest path between a certain number of cities is to be found without passing one city twice. Using conventional methods, like the backtracking algorithm, this problem can't be solved on any computer in time for a number of cities greater than about 25. The presented neural net is able to solve this problem for even 50 or more cities. And all that in three dimensions! The result may not definitely be the optimal solution but it's always a good approximation if all parameters have been properly set.
The structure of the feature map may be changed in the Parameters window. It can either be one-dimensional (displayed as a black line) or two-dimensional (displayed as a black grid). (see Sample Applet Controls for a detailed description of all changeable parameters)
If you want to know, how the learning algorithm of a Kohonen Feature Map works, take a closer look on the "Selforganization" section in neural net overview/The learning process.