Systematic behavior of fragments in Bayesian neural network models for projectile fragmentation reactions

The recently proposed Bayesian neural network (BNN) model was adopted to investigate the systematic behaviors of fragments cross sections produced in projectile fragmentation reactions. Mainly two phenomena are studied, i.e., (1) The scaling behavior in the difference between mass excess of mirror nuclei based on the binding energies of proton-rich fragments determined using the BNN-predicted cross sections for the 345A MeV 78Kr+9Be reaction, which shows that the scaling phenomenon exists up to neutron excess |I| = |N - Z| of 5; (2) the isobaric yield ratio distributions for mirror fragments [IYR(m)] produced in a series of projectile fragmentation reactions with projectiles of isotopes, isobars, and those that have the same neutron-skin thickness at different levels. The IYR(m) distributions predicted by the BNN model reflect the systematic evolution with neutron-skin thickness of projectile nuclei.