Title :
Improving the efficiency of Sammon´s nonlinear mapping by using clustering archetypes
Author_Institution :
Royal Signals & Radar Establishment, Great Malvern, UK
Abstract :
Sammon´s nonlinear mapping technique is discussed and certain inefficiencies are highlighted. It is shown how the algorithm can be streamlined by using appropriately chosen cluster archetypes. This results in reductions in storage requirements and in execution time, the latter by at least an order of magnitude.
Keywords :
computerised pattern recognition; data reduction and analysis; cluster analysis; cluster archetypes; data reduction; nonlinear mapping technique; pattern recognition;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:19780539
