CLASSIFICATION OF SPECTRA OF NORMAL STARS OF THE MAIN SEQUENCE OF THE NEURAL NETWORK, DEEP LEARNING

Abstract

ence stars and gives a brief description of the research method. The сonvolutional
Neural Network (CNN) was tested in a classification on two well-known data sets and achieved
positive results even with a very small number of teaching samples in the classroom. On the basis of
the obtained results, it can be argued that the CNN deep learning model is an effective tool in the
classification problems of the main sequence stars. The use of a large number of training samples
increases computational accuracy regardless of the complexity of the waveform. The use of
artificial neural networks shows its advantages precisely in processing tasks, classifications,
identifying large data sets of complex signals, such as spectra of main sequence stars, etc.. Effective
processing of large data sets, using a smaller resource base in a short processing time, is an urgent
task of modern research, including the processing of astrophysical signals.