ИСТИНА

The paper presents the results of comparative analysis of input data compression methods for simultaneous neural network based solution of the problems of determination of types and concentrations of dissolved inorganic salts in multi-component water solutions by Raman spectra. It is demonstrated that use of uniform aggregation of input features allows decreasing the error of determination of partial concentrations of components by 16% on the average, with simultaneous reduction of the requirements to the resolution of spectral equipment. Use of non-uniform aggregation allows further error reducing for determination of partial concentrations of salts with complex anions.