Neural network classification of discontinuities in space plasma parameters

The artificial neural network technique is applied to dividing discontinuities in space plasma and magnetic field parameters into classes corresponding to known types of magnetohydrodynamic discontinuities. Parameter discontinuities registered on the WIND spacecraft between 1996 and 1999 are classified using a network of the Kohonen Layer type. An algorithm for determining the orientation of discontinuity surfaces on the basis of one-dimensional observations of solar wind parameter discontinuities on board spacecraft is proposed.