Title :
A new quantitative measure of topology preservation in Kohonen´s feature maps
Author :
Villmann, Th. ; Der, R. ; Martinetz, Th
Author_Institution :
Inst. of Inf., Leipzig Univ., Germany
fDate :
27 Jun-2 Jul 1994
Abstract :
In this paper we give a new approach for quantifying topology preservation using explicitly the structure of the data manifold. It can be applied to linear and nonlinear data manifolds M. Further, this method allows one to quantify the range of folds. Our approach employs what we call the topographic function, which is defined based on the so called masked Voronoi polyhedra introduced by Martinetz (1993) for defining neighbourhood and topology preservation of feature maps
Keywords :
learning (artificial intelligence); self-organising feature maps; topology; Kohonen´s feature maps; data manifold; masked Voronoi polyhedra; quantitative measure; topographic function; topology preservation; Euclidean distance; Graphics; Informatics; Lattices; Particle measurements; Research and development; Robots; Topology;
Conference_Titel :
Neural Networks, 1994. IEEE World Congress on Computational Intelligence., 1994 IEEE International Conference on
Conference_Location :
Orlando, FL
Print_ISBN :
0-7803-1901-X
DOI :
10.1109/ICNN.1994.374251