Method for informative feature selection on a model IR spectra
Denis A. Vrazhnov, 1,2 Tatyana V. Kabanova, 3, Tatyana E. Malakhova, 3, Alexey V. Borisov, 2, Yuri V. Kistenev, 2
1 V.E. Zuev Institute of Atmospheric optics SBRAS, Tomsk, Russia,
2 Tomsk State University, Laboratory of biophotonics, Tomsk, Russia
3 Tomsk State University, Institute of Applied Mathematics and Computer Science, Tomsk, Russia
Abstract
The paper presents an algorithm based on low order statistics for the informative feature extraction for IR spectra. The proposed method was tested on the HITRAN modeled IR spectra. Supervised and unsupervised machine learning methods were applied to select the most informative frequencies and test the processed data separability. 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)
File with abstract
Speaker
Denis Aleksandrovich Vrazhnov
Tomsk State University
Russia
Discussion
Ask question
