Application of Neural Networks Optimized by Genetic Algorithms to Higgs Bosson Search

0404660 - UIVT-O 20020054 RIV GB eng C - Konferenční příspěvek (zahraniční konf.)
Hakl, František - Hlaváček, M. - Kalous, R.
Application of Neural Networks Optimized by Genetic Algorithms to Higgs Bosson Search.
Advanced Statistical Techniques in Particle Physics. Durham: University of Durham, 2002 - (Whalley, M.; Lyons, L.), s. 124-129
[Conference on Advanced Statistical Techniques in Particle Physics. Durham (GB), 18.03.2002-22.03.2002]
Grant CEP: GA MPO RP-4210/69/97
Výzkumný záměr: AV0Z1030915
Klíčová slova: neural network * genetic algorithms * global optimization * Higgs boson search * Mbb distribution
Kód oboru RIV: BX

Neural network (NN) model with a special type of data flow is used to separate ttjj background from H-->bb events. A genetic algorithm principles are used to tune parameters of NN in order to optimize the final NN performance. Our results show that NN filters can be used to emphasize shape of M_bb distribution for events accepted by filter. This improvement can be used as a criterion of existence of Higgs boson decay in considered data.
Trvalý link: http://hdl.handle.net/11104/0124900