Bereiche | Tage | Auswahl | Suche | Downloads | Hilfe

HK: Hadronen und Kerne

HK 56: Postersitzung

HK 56.79: Poster

Mittwoch, 18. März 1998, 17:00–19:00, Foyer

Application of Self Organizing Maps in the EDDA data-analysis — •M. Busch — Institut f"ur Strahlen- und Kernphysik, Universit"at Bonn, Nu"sallee 14-16, D-53115 Bonn

The EDDA-experiment aims to measure excitation functions, angular distributions and spin-correlation parameters of elastic proton-proton scattering in the energy range from 500 to 2500MeV. Self Organizing Maps are used in data-analysis to improve the separation of elastic events from background reactions, especially for data aquired with CH₂-fiber targets.

A SelfOrganizing Map projects sets of high dimensional experimental input data onto a discrete low dimensional structure (chain, grid) in such a way, that events with similar features are projected onto neighboring elements of the discrete structure. This property of topology-preserving can be used to isolate different event-classes of physical reactions by use of the complete experimental information provided by the EDDA detector without resort to Monte-Carlo Simulations.