Weighted partial least squares based on the error and variance of the recovery rate in calibration set

doi:10.1016/j.saa.2017.04.029

	Weighted partial least squares based on the error and variance of the recovery rate in calibration set
	Yu, Shaohui 1; Xiao, Xue 2; Ding, Hong 1; Xu, Ge 1; Li, Haixia 1; Liu, Jing 3
	2017-08-05
发表期刊	SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY
摘要	The quantitative analysis is very difficult for the emission-excitation fluorescence spectroscopy of multi-component mixtures whose fluorescence peaks are serious overlapping. As an effective method for the quantitative analysis, partial least squares can extract the latent variables from both the independent variables and the dependent variables, so it can model for multiple correlations between variables. However, there are some factors that usually affect the prediction results of partial least"squares, such as the noise, the distribution and amount of the samples in calibration set etc. This work focuses on the problems in the calibration set that are mentioned above. Firstly, the outliers in the calibration set are removed by leave-one-out cross-validation. Then, according to two different prediction requirements, the EWPLS method and the VWPLS method are proposed. The independent variables and dependent variables are weighted in the EWPLS method by the maximum error of the recovery rate and weighted in the VWPLS method by the maximum variance of the recovery rate. Three organic matters with serious overlapping excitation-emission fluorescence spectroscopy are selected for the experiments. The step adjustment parameter, the iteration number and the sample amount in the calibration set are discussed. The results show the EWPLS method and the VWPLS method are superior to the PIS method especially for the case of small samples in the calibration set. (C) 2017 Elsevier B.V. All rights reserved.
文章类型	Article
关键词	Leave-one-out Cross-validation Partial Least Squares Gauss Weight Excitation-emission Fluorescence Spectroscopy
WOS标题词	Science & Technology ; Technology
DOI	10.1016/j.saa.2017.04.029
关键词[WOS]	NEAR-INFRARED SPECTROSCOPY ; PLS-REGRESSION ; ALGORITHMS ; ROBUST
收录类别	SCI
语种	英语
项目资助者	National Natural Science Foundation of China(61308063 ; National Natural Science Foundation of China(61308063 ; National Natural Science Foundation of China(61308063 ; National Natural Science Foundation of China(61308063 ; National Natural Science Foundation of China(61308063 ; National Natural Science Foundation of China(61308063 ; National Natural Science Foundation of China(61308063 ; National Natural Science Foundation of China(61308063 ; Projects of International Cooperation and Exchanges NSFC(61491240110) ; Projects of International Cooperation and Exchanges NSFC(61491240110) ; Projects of International Cooperation and Exchanges NSFC(61491240110) ; Projects of International Cooperation and Exchanges NSFC(61491240110) ; Projects of International Cooperation and Exchanges NSFC(61491240110) ; Projects of International Cooperation and Exchanges NSFC(61491240110) ; Projects of International Cooperation and Exchanges NSFC(61491240110) ; Projects of International Cooperation and Exchanges NSFC(61491240110) ; Anhui Provincial Natural Science Foundation(1308085QF111) ; Anhui Provincial Natural Science Foundation(1308085QF111) ; Anhui Provincial Natural Science Foundation(1308085QF111) ; Anhui Provincial Natural Science Foundation(1308085QF111) ; Anhui Provincial Natural Science Foundation(1308085QF111) ; Anhui Provincial Natural Science Foundation(1308085QF111) ; Anhui Provincial Natural Science Foundation(1308085QF111) ; Anhui Provincial Natural Science Foundation(1308085QF111) ; Student Innovation Training Program(201614098087 ; Student Innovation Training Program(201614098087 ; Student Innovation Training Program(201614098087 ; Student Innovation Training Program(201614098087 ; Student Innovation Training Program(201614098087 ; Student Innovation Training Program(201614098087 ; Student Innovation Training Program(201614098087 ; Student Innovation Training Program(201614098087 ; 61378041) ; 61378041) ; 61378041) ; 61378041) ; 61378041) ; 61378041) ; 61378041) ; 61378041) ; 201614098013) ; 201614098013) ; 201614098013) ; 201614098013) ; 201614098013) ; 201614098013) ; 201614098013) ; 201614098013)
WOS研究方向	Spectroscopy
WOS类目	Spectroscopy
WOS记录号	WOS:000403130900019
引用统计	被引频次：3[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://ir.hfcas.ac.cn:8080/handle/334002/31907
专题	中科院安徽光学精密机械研究所
作者单位	1.Hefei Normal Univ, Sch Math & Stat, Hefei 230061, Peoples R China 2.Chinese Acad Sci, Anhui Inst Opt & Fine Mech, Key Lab Environm Opt &Technol, Hefei 230031, Peoples R China 3.Chinese Acad Sci, Inst Tech Biol & Agr Engn, Key Lab Ion Beam Bioengn, Hefei 230031, Peoples R China
推荐引用方式 GB/T 7714	Yu, Shaohui,Xiao, Xue,Ding, Hong,et al. Weighted partial least squares based on the error and variance of the recovery rate in calibration set[J]. SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY,2017,183(无):138-143.
APA	Yu, Shaohui,Xiao, Xue,Ding, Hong,Xu, Ge,Li, Haixia,&Liu, Jing.(2017).Weighted partial least squares based on the error and variance of the recovery rate in calibration set.SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY,183(无),138-143.
MLA	Yu, Shaohui,et al."Weighted partial least squares based on the error and variance of the recovery rate in calibration set".SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY 183.无(2017):138-143.

条目包含的文件		下载所有文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
Weighted partial lea（958KB）	期刊论文	作者接受稿	开放获取	CC BY-NC-SA	浏览下载