AUTOMATIC-ANALYSIS OF JET CHARGE-EXCHANGE SPECTRA USING NEURAL NETWORKS

The analysis of charge exchange recombination spectra represents a ver y challenging problem due to the presence of many overlapping spectral lines. Conventional approaches are based on iterative least-squares o ptimization and suffer from the two difficulties of low speed and the need for a good initial approximation to the solution. This latter pro blem necessitates considerable human supervision of the analysis proce dure. In this paper we show how neural network techniques allow charge exchange data to be analysed very rapidly, to give an approximate sol ution without the need for supervision. The network approach is well s uited to the fast intershot analysis of large volumes of data, and can readily be implemented in dedicated hardware for real-time applicatio ns. The neural network can also be used to provide the initial guess f or the standard least-squares algorithm when high accuracy is required .