Decoding functions for Kohonen maps
The Kohonen maps (KM) connectionist model is a technique for unsupervised data analysis which maps a feature space onto a low-dimensional lattice of reference vectors while attempting to preserve some topological properties of the original sample distribution. This paper addresses the issue of defining a suitable decoding function. From a trained KM, for which the reference vectors and number of training patterns that win on each cell are known, the mean value of the training samples belonging to the receptive field of any cell are estimated.
Bibliographic Reference: Paper presented: European Symposium on Artificial Neural Networks ESANN '94, Bruxelles (BE), April 20-22, 1994
Availability: Available from (1) as Paper EN 38157 ORA
Record Number: 199410385 / Last updated on: 1994-11-28
Original language: en
Available languages: en