Adaptive Hierarchical Incremental Grid Growing:
   An architecture for high-dimensional data visualization


Dieter Merkl(1,2), Shao Hui He(3),  Michael Dittenbach(2), and Andreas Rauber(3)

(1) Institut fuer Rechnergestuetzte Automation, 
    Technische Universitaet Wien
    http://www.rise.tuwien.ac.at/

(2) Electronic Commerce Competence Center--EC3, Wien
    http://www.ec3.at/

(3) Institut fuer Softwaretechnik und Interaktive Systeme, 
    Technische Universitaet Wien
    http://www.ifs.tuwien.ac.at/




Based on the principles of the self-organizing map, we have designed a novel
neural network model with a highly adaptive hierarchically structured 
architecture, the adaptive hierarchical incremental grid growing.
This feature allows it to capture the unknown data topology in terms 
of hierarchical relationships and cluster structures in a highly accurate way.
In particular, unevenly distributed real-world data is represented in 
a suitable network structure according to its specific requirements 
during the unsupervised training process.
The resulting three-dimensional arrangement of mutually independent maps 
reveals a precise view of the inherent topology of the data set.

---------------

In:  Proceedings of the Workshop on Self-Organizing Maps (WSOM 2003), pp. 293-298, September 11-14 2003, Kitakyushu, Japan.