Advantages of neural networks for deriving an electrons spectrum from depth-charge curve

The work considers the artificial intelligence methods for solving ill-posed problem of reconstruction of the electrons spectrum from the distribution of the charge deposition. In order to perform comparison of different neural networks types, method of least squares and Tikhonov regularization the existing computer models of physical processes are used. The results deviation, variance and probability of obtaining the results without physical meaning are used for effectiveness criteria. Results of the work shows the advantages of the radial basis neural networks.