Application of deep learning and genetic algorithms to the solution of absorbance spectroscopy component analysis of gas mixtures problem
Vladimir V. Prischepa,1 Victor E. Skiba,1 Denis A. Vrazhnov,1,2 1 Tomsk State University, Tomsk, Russia 2 V.E. Zuev Institute of Atmospheric Optics, Siberian Branch of the RAS, Tomsk, Russia
Abstract
A well-known method of algebraic reconstruction applied to the component analysis task of gas mixtures may give poor results because of noisy input data. A brute force method for choosing the most reliable part of the data cannot be used due to large dimensionality. We propose a sequential method based on an evolutionary algorithm for solving component analysis problem. The results are compared with a deep learning method, also developed by the authors.
The research was carried out with the support of a grant under the Decree of the Government of the Russian Federation No. 220 of 09 April 2010 (Agreement No. 075-15-2021-615 of 04 June 2021).
Speaker
Vladimir Prischepa
Tomsk State University
Russia
Report
File with report
Discussion
Ask question
