The problem of smoothing of radiation spectra is one of the issues of preliminary data processing. It is solving in complex with the problems of localization of resonance lines in spectra and of approximation of background component. To smooth spectra the convolution with a Gaussian filter of optimal width are used.

Based on X-ray, gamma-ray, and charged-particle measurements with several instruments onboard the Coronas-F satellite and on ACE and GOES experimental data presented on the Internet, we investigate the parameters of the solar flare of November 4, 2001, and the energetic-particle fluxes produced by it in circumterrestrial space. The increase in relativistic-electron fluxes for about 1.5 days points to a moving source (shock front). The structure of the energetic-particles fluxes in the second half of November 5, 2001, can be explained by the passage of the coronal mass ejection that was ejected on November 1, 2001, and that interacted with the shock wave from the flare of November 4, 2001.

We consider mathematical and information problems of elemental analysis of a substance and of radiation based on radiological methods of measurement and control. We propose a technology for solving problems of the elemental analysis at an algorithmic level. Certain results of operation of automatic data processing systems are presented.

Mathematical models of radiation spectra are considered. The tasks and algorithms for the main data processing stages are described: decomposition of multiplets into singlets, approximation and precision correction of the calibration coefficients, determination of element (nuclide) content, estimation of the linear and nonlinear parameters of the apparatus weighting function. Analysis of alpha-radiation spectra is used as an example to illustrate the various analytical issues in the presence of strong asymmetry of the apparatus weighting function. The solution of element analysis problems exploits the bilinear relationship of the linear parameters in the approximating model and the quasi=exponential dependence on the nonlinear parameters. The procedure successively applies groups of computer algorithms combined with simulation using empirical databases.

The article examines the mathematical model of recorded data and issues of sample element analysis by radionuclide methods. The proposed element-analysis technology is software-implemented in an intelligent interactive-graphical environment, which provides opportunities for active interaction with data banks and inclusion of expert support. This environment makes it possible to apply complex data-processing schemes coded in a special-purpose language and to adapt processing models and procedures to the specific features of data streams and to properties of the technological process or experiment.