Identification of muons in the H1 LAr-calorimeter Neural networks vs. fuzzy elements

We present and compare the application of different algorithms for the identification of low momentum muons (p ≈ 1 GeV) in the LAr calorimeter of the H1 Experiment at HERA. The first method for muon identification applies conventional cuts on characteristic variables of the signature in the LAr calorimeter, the second one uses a standard feed forward neural network. The third one which is used in the standard H1 reconstruction software combines the first method with elements of fuzzy logic.