Waveband selection of Infrared Spectrum Based on Sine Cosine Algorithm

Infrared spectrum analysis is the main technical mean for qualitative and quantitative analysis of gase mixture. In the process of spectral analysis, too many dimensions of infrared spectral data and too much irrelevant information will have adverse effects on the modeling effect. Therefore, waveban...

Full description

Saved in:
Bibliographic Details
Published in2021 China Automation Congress (CAC) pp. 483 - 487
Main Authors Li, Yujun, Yang, Zhi, Jiao, Shangbin, Zhang, Qing, Wang, Qing, Du, Jinghang
Format Conference Proceeding
LanguageEnglish
Published IEEE 22.10.2021
Subjects
Analytical models
Data mining
Data models
Gases
Interference
Methane
Near Infrared Spectroscopy
Sine Cosine Algorithm
Statistical analysis
Waveband extraction
Online AccessGet full text

Cover

More Information
Summary:Infrared spectrum analysis is the main technical mean for qualitative and quantitative analysis of gase mixture. In the process of spectral analysis, too many dimensions of infrared spectral data and too much irrelevant information will have adverse effects on the modeling effect. Therefore, waveband selection is very important to improve the accuracy of model analysis. In this paper, infrared spectrum data which obtained from 938 gas mixture samples is prepared for component gases quantiative analysis, which the concentration ranges from 1% to 10% (methane), 1% to 15% (ethane), and 1% to 10% (propane). In order to improve the accuracy of the component gas quantiative analysis model, sine cosine algorithm (SCA) is proposed to select the waveband of the infrared spectrum data of gas mixture. Experiments results show that the average MRE of the three component gas analysis models deduce from 47.98% to 23.54%, and the average dimentions of the selected waveband spenctrum data deduced from 1866 to 270. The model precision and calculation efficiency are greatly improved.
ISSN:2688-0938
DOI:10.1109/CAC53003.2021.9728104