Machine learning based event classification for the energy-differential measurement of the <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline" id="d1e721" altimg="si4.svg"><mml:msup><mml:mrow /><mml:mrow><mml:mtext>nat</mml:mtext></mml:mrow></mml:msup></mml:math>C(n,p) and <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline" id="d1e732" altimg="si4.svg"><mml:msup><mml:mrow /><mml:mrow><mml:mtext>nat</mml:mtext></mml:mrow></mml:msup></mml:math>C(n,d) …

The paper explores the feasibility of using machine learning techniques, in particular neural networks, for classification experimental data from joint $^\text{nat}$C(n,p) and $^\text{nat}$C(n,d) reaction cross section measurement neutron time flight facility n_TOF at CERN. Each relevant $\Delta E$-$E$ pair strips two segmented silicon telescopes is treated separately afforded its own dedicated network. An important part procedure a careful preparation training datasets, based on raw Geant4 simulations. Instead these we divide 3-parameter space into discrete voxels, classify each voxel according to particle/reaction type submit voxels procedure. capabilities structurally optimized trained networks are found be superior those manually selected cuts.