The feasibility of using neural networks to obtain cross sections from electron swarm data

The use of an artificial neural network as an optimization technique for treating the inverse problem of obtaining electron collision cross section from electron transport data is explored in which electron-impact cross sections from measured drift velocities, characteristic energies, and other swarm data are obtained. Momentum transfer cross sections obtained for a model problem and for xenon using a neural network are presented.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>