The Spherical Self-Organizing Feature Maps is a contemporary technique for data clustering and visualization.
The main advantages they offer are the following:
1. Smooth training
2. Implementation in arbitrary dimension without additional computational cost.
3. Data visualization in arbitrary dimensions
This toolbox contains a set of functions and a GUI which can be used to create glyphs from spherical Self-Organizing Feature Maps (SOFMs). It is freely distributable for educational and research purposes.
Example:
% loads the data to be visualized
load henon-1024-4.mat
%loads the S-SOFM structure
load c4-24.mat
%trains the S-SOFM
[w,g,r]=trainGlyph(P,X,C,0.5,20,'plot');
% plots the S-SOFM
glyph(X,r); |
