Title :
Monitoring the formation of kernel-based topographic maps
Author :
Van Hulle, Marc M.
Author_Institution :
Lab. voor Neuro- en Psychofysiologie, Katholieke Univ., Leuven, Belgium
Abstract :
Topographic maps have attracted the attention of the data mining community since they can be used for representing and visualizing multidimensional data. For applications like these, it is crucial that the maps are free of topological defects. We introduce a new algorithm for monitoring the degree of topology preservation of kernel-based maps during learning. The algorithm is applied to a synthetic example in this article, and a large, real-world example in our companion article
Keywords :
data mining; data visualisation; learning (artificial intelligence); self-organising feature maps; topology; data mining; kernel-based topographic maps; learning; multidimensional data visualization; neural network; self organizing maps; topological defects; topology preservation; Data mining; Data visualization; Electronic mail; Laboratories; Lattices; Monitoring; Multidimensional systems; Neurons; Psychology; Topology;
Conference_Titel :
Neural Networks for Signal Processing X, 2000. Proceedings of the 2000 IEEE Signal Processing Society Workshop
Conference_Location :
Sydney, NSW
Print_ISBN :
0-7803-6278-0
DOI :
10.1109/NNSP.2000.889415
