Abstract
A model combining UV–visible (UV-Vis) spectroscopy and support vector regression (SVR) for the quantitative detection of thiamethoxam in tea is proposed. First, each original UV-Vis spectrum in the sample set is decomposed into some intrinsic mode functions (IMFs) and a residual via ensemble empirical mode decomposition. Next, the decomposed IMFs are reconstructed into high-frequency and low-frequency matrices, and the residuals are combined into a trend matrix. Then, the SVR is used to build regression sub-models between each matrix and the content of thiamethoxam in tea. Finally, the combination model is established by a weighted average of the sub-models. The prediction results are compared with SVR and SVR coupled with several preprocessing methods, and the results demonstrate the superiority of the proposed approach in the quantitative detection of thiamethoxam in tea.
© 2022 Optica Publishing Group
Full Article | PDF ArticleMore Like This
Rendong Ji, Zhezhen Jiang, Xiaoyan Wang, Yue Han, Haiyi Bian, Yudong Yang, Liyun Zhuang, and Yulin Zhang
Appl. Opt. 61(12) 3455-3462 (2022)
Wu Pan, Huiyun Xiao, Haizhu Li, Yongrui Li, Bin Zhang, Bowen Liu, and Longliang Yang
Appl. Opt. 61(33) 9904-9910 (2022)
Wei Wang and Xiyuan Chen
Appl. Opt. 57(28) 8443-8450 (2018)