Abstract
<jats:p>This paper explores the possibility of using machine learning methods for analyzing observations from the “Solntse-Terahertz” scientific equipment, developed at the Lebedev Physical Institute for installation on the Russian segment of the ISS. The scientific equipment consists of eight detectors, with target frequencies ranging from 0.4 to 12.0 THz. One of the primary goals of the experiment is to study solar flares whose spectra in this range often have a U-shaped form. The primary focus in determining the spectral parameters is on identifying spectral indices of the decaying and rising parts of the spectrum, as well as the position of the turnover point. The algorithms were trained using model data on the intensity of radiation passing through optical paths of the instrument. The data was obtained by numerical integration methods. The analysis has shown that the Stacking algorithm demonstrates the highest accuracy in determining the spectral parameters and can be integrated into the data processing system for future experiment on the ISS, enabling the automatic preliminary restoration of solar flare spectrum parameters.</jats:p>