A feedforward multilayered neural network has been trained to "recognize" true V0's in the presence of a large combinatoric background using simulated data for 2 GeV/nucleon Ni + Cu interactions. The resulting neural network filter has been applied to actual data from the EOS TPC experiment. An enhancement of signal to background over more traditional selection mechanisms has been observed.
Cite
@article{arxiv.nucl-ex/9612001,
title = {Separating True V0's from Combinatoric Background with a Neural Network},
author = {Marvin Justice},
journal= {arXiv preprint arXiv:nucl-ex/9612001},
year = {2016}
}
Comments
9 pages LaTeX using elsart.sty and psfig, 5 ps figures. Submitted to NIM