Qualitative and Quantitative Identification of Components in Mixture by Terahertz Spectroscopy
Authors:
- Yan PengYPYan Peng
- Chenjun ShiCSChenjun Shi
- Mingqian XuMXMingqian Xu
- Tianyi KouTKTianyi Kou
- Xu WuXWXu Wu
- Bin SongBSBin Song
- Hongyun MaHMHongyun Ma
- Shiwei GuoSGShiwei Guo
- Lizhuang LiuLLLizhuang Liu
- Yiming ZhuYiming Zhu
Open Access
- 29 August 2018
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Terahertz Science and Technology
- Vol. 8 (6) , 696-701
Abstract
For the mixture components analysis, the typical methods include multiple linear regression, partial linear squares and artificial neural network. However, they all need lots of samples and time to improve the recognition accuracy. Here, in this paper, based on the data obtained from terahertz spectroscopy, we proposed an identification method with less sample requirements and less calculation time but higher accuracy. Basing on the wavelet transform, baseline elimination, support vector regression and loop iteration of samples, the specific substance in mixture can be identified effectively. For example, seven substances existing in brain glioma are chosen as the components of mixture, where the key substances used for glioma diagnostic are set as the target substances, and the spectra of mixtures with different mix proportions are served as training data. The average correlation coefficient of identification achieves 99.135% and root mean square error (RMSE) is 0.40%. These results have profound meaning for the eventual practical application of exact qualitative and quantitative identification of components in mixtures.Keywords
Funding Information
- National Program on Key Basic Research Project of China (2014CB339806)
- Major National Development Project of Scientific Instrument and Equipment (2017YFF0106300, 2016YFF0100503)
- National Natural Science Foundation of China (61771314, 61722111)
- Shanghai Pujiang Program (16PJD033)
- Shanghai Rising-Star Program (17QA1402500)
- Shuguang Program (17SG45)
- Shanghai Youth Talent Support Program and Young Yangtse River Scholar (Q2016212)
This publication has 23 references indexed in Scilit:
- Quantitative Analysis of Mixtures Using Terahertz Time-Domain Spectroscopy and Different PLS AlgorithmsAdvanced Materials Research, 2013
- Terahertz spectroscopy and imaging – Modern techniques and applications [Laser Photon. Rev. 5, No. 124–166 (2011)]Laser & Photonics Reviews, 2012
- Using wavelet transform and multi-class least square support vector machine in multi-spectral imaging classification of Chinese famous teaExpert Systems with Applications, 2011
- Wavelet-Based Cascaded Adaptive Filter for Removing Baseline Drift in Pulse WaveformsIEEE Transactions on Biomedical Engineering, 2005
- Far-infrared spectra of amino acidsChemical Physics, 2005
- Noise and background removal in Raman spectra of ancient pigments using wavelet transformJournal of Raman Spectroscopy, 2005
- Terahertz technologyIEEE Transactions on Microwave Theory and Techniques, 2002
- Support-vector networksMachine Learning, 1995
- Parallelizing Mallat algorithm for 2-D wavelet transformsInformation Processing Letters, 1993
- Single-layer learning revisited: a stepwise procedure for building and training a neural network1990