不同特征选择方法结合激光诱导击穿光谱量化不锈钢镍、钛和铬元素含量

doi:10.3964/j.issn.1000-0593(2024)12-3339-08

摘要
参考文献
相关文章 (15)

全文: PDF (5531 KB)
输出: BibTeX | EndNote (RIS)

摘要：激光诱导击穿光谱（LIBS）技术作为物质成分分析的一种有效工具，具有广泛的应用价值，但由于LIBS可重复性差以及受基体效应和自吸收效应影响等，导致光谱数据中包含大量对定量分析无用的冗余特征。为了克服使用原始全光谱数据作为模型输入时导致预测精度难以提高的不足，利用两种特征工程技术（最小绝对收缩和选择算子回归 LASSO和顺序后向选择SBS）结合机器学习实现对不锈钢样品中镍（Ni）、钛（Ti）和铬（Cr）三种元素的量化分析。研究使用购自钢研纳克检测技术股份有限公司的7种元素含量相异的不锈钢样品为研究对象，实验得到70个LIBS光谱，比较了四种不同的数据预处理方法包括最大最小归一化（MMN）、标准正态变换（SNV）、Savitzky-Golay平滑滤波（SG）以及内标法（IS），以均方根误差（RSME）检测预处理结果，最终选择Savitzky-Golay平滑滤波进行光谱预处理，在使用LASSO算法和SBS算法选择特征时，针对不同量化元素独立进行有效变量的选取，然后使用全光谱、LASSO选择特性、SBS选择特征三种不同的特征组合，作为模型的输入，为了验证特征选择方法的有效性，在偏最小二乘法（PLS）、支持向量机（SVM）两种不同机器学习模型中进行对比。使用平均相对误差（ARE）和相对标准偏差（RSD）来评估不同模型的性能。结果显示，两种特征选择方法选择的模型输入相比全光谱输入在不同的机器学习模型中都显示出更加优秀的预测精度和稳定性，其中LASSO-PLS模型在Ni、Ti、Cr元素的量化分析上得到最佳的预测准确度，ARE分别为3.50%、2.66%、0.93%，RSD分别为4.55%、5.23%、2.04%。因此，本文提出的LIBS结合LASSO和SBS算法能够准确稳定地对不锈钢中的Ni、Ti、Cr元素进行量化分析，对进一步发掘LIBS结合机器学习在不锈钢元素量化分析场景提供参考。

关键词：激光诱导击穿光谱；不锈钢元素含量；特征选择；定量分析

Abstract：Laser-induced breakdown spectroscopy (LIBS) technology, as an effective tool for material composition analysis, has broad application value. However, due to the poor repeatability of LIBS and the influence of matrix and self-absorption effects, spectral data contains a large number of redundant features that are useless for quantitative analysis. To overcome the difficulty in improving prediction accuracy when using raw full spectrum data as model input, two feature engineering techniques (minimum absolute shrinkage and selection operator regression LASSO and sequential backward selection SBS) were combined with machine learning to achieve a quantitative analysis of nickel (Ni), titanium (Ti), and chromium (Cr) in stainless steel samples. This study used seven stainless steel samples with different element contents purchased from Steel Research Nanogram Testing Technology Co., Ltd. as the research objects. Seventy LIBS spectra were obtained, and four different data preprocessing methods were compared, including Maximum Minimum Normalization (MMN), Standard Normal Variation (SNV), Savitzky Golay Smooth Filtering (SG), and Internal Standard Method (IS). The preprocessing results were detected using Root Mean Square Error (RSME). Finally, Savitzky Golay smoothing filtering was chosen for spectral preprocessing. Effective variables were independently selected for different quantization elements when selecting features using LASSO and SBS algorithms. Then, three different feature combinations, namely full spectrum, LASSO selection feature, and SBS selection feature were used as inputs to the model. To verify the effectiveness of the feature selection method, partial least squares (PLS) Compare two different machine learning models using a Support Vector Machine (SVM). Evaluate the performance of different models using Average Relative Error (ARE) and Relative Standard Deviation (RSD). The results showed that the model inputs selected by the two feature selection methods showed better prediction accuracy and stability compared to full-spectrum inputs in different machine learning models. Among them, the LASSO-PLS model achieved the best prediction accuracy in the quantitative analysis of Ni, Ti, and Cr elements, with ARE of 3.50%, 2.66%, and 0.93%, and RSD of 4.55%, 5.23%, and 2.04%, respectively. Therefore, the LIBS combined with LASSO and SBS algorithms proposed in this article can accurately and stably quantify the Ni, Ti, and Cr elements in stainless steel, providing a reference for further exploring the application of LIBS combined with machine learning in stainless steel element quantification analysis scenarios.

Key words：Laser-induced breakdown spectroscopy; Stainless steel element content; Sequential backward selection; Quantitative analysis

收稿日期: 2023-11-11 修订日期: 2024-03-15

中图分类号:

O433.4

基金资助: 湖北省重点研发计划项目（2023BAB063），国家重点研发计划项目（2024YFF0728300），国家重大科研仪器研制项目（41927803）资助

通讯作者: 吕涛 E-mail: lvtaohn@126.com

作者简介: 吴卓，2000年生，中国地质大学（武汉）自动化学院硕士研究生 e-mail: 1583855135@qq.com

引用本文:

吴卓，苏晓慧，范博文，朱惠会，张宇博，方彬，王一帆，吕涛. 不同特征选择方法结合激光诱导击穿光谱量化不锈钢镍、钛和铬元素含量[J]. 光谱学与光谱分析, 2024, 44(12): 3339-3346.
WU Zhuo, SU Xiao-hui, FAN Bo-wen, ZHU Hui-hui, ZHANG Yu-bo, FANG Bin, WANG Yi-fan, LÜ Tao. Different Feature Selection Methods Combined With Laser-Induced Breakdown Spectroscopy Were Used to Quantify the Contents of Nickel, Titanium and Chromium in Stainless Steel. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2024, 44(12): 3339-3346.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2024)12-3339-08 或 https://www.gpxygpfx.com/CN/Y2024/V44/I12/3339

[1] Costa V C, Augusto A S, Castro J P, et al. Quimica Nova, 2019, 42(5): 527.
[2] Chahrour O, Malone J. Protein and Peptide Letters, 2017, 24(3): 253.
[3] Talari A C S, Movasaghi Z, Rehman S, et al. Applied Spectroscopy Reviews, 2015, 50(1): 46.
[4] Porep J U, Kammerer D R, Carle R. Trends in Food Science & Technology, 2015, 46(2): 211.
[5] Li F, Ge L Q, Tang Z Y, et al. Applied Spectroscopy Reviews, 2020, 55(4): 263.
[6] Terán E J, Montes M L, Rodríguez C, et al. Microchemical Journal, 2019, 144: 159.
[7] Markiewicz-Keszycka M, Cama-Moncunill X, Casado-Gavalda M P, et al. Trends in Food Science & Technology, 2017, 65: 80.
[8] Agresti J, Indelicato C, Perotti M, et al. Molecules, 2022, 27(6): 1813.
[9] Millar S, Kruschwitz S, Wilsch G. Cement and Concrete Research, 2019, 117: 16.
[10] SUN Lan-xiang, WANG Wei, ZHANG Peng, et al(孙兰香, 汪为, 张鹏, 等). Metallurgical Analysis(冶金分析), 2021, 41(12): 58.
[11] Zhang Q H, Liu Y Z. Atomic Spectroscopy, 2022, 43(2): 174.
[12] LIU Yu-duo, SHEN Wei-hua, ZHU Zhi-qing, et al(刘雨朵, 沈卫华, 朱志庆, 等). Corrosion and Protection(腐蚀与防护), 2022, 43(12): 1.
[13] XU Xiu-qing, WANG Wei, CHEN Zhi-teng, et al(徐秀清, 王玮, 陈之腾, 等). Materials Reports(材料导报), 2022, 36(14): 21030183.
[14] LI Long-bo, LI Zheng-xian, LIU Lin-tao, et al(李龙博, 李争显, 刘林涛, 等). Rare Metal Materials and Engineering(稀有金属材料与工程), 2021, 50(5): 1743.
[15] Mamonova A A, Baglyuk G A, Kurovskii V Y. Physics of Metals and Metallography, 2015, 116(6): 562.
[16] Barua A, Ahmed M U, Begum S. IEEE Access, 2023, 11: 14804.
[17] ZHAO Wen-ya, MIN Hong, LIU Shu, et al(赵文雅, 闵红, 刘曙, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2021, 41(7): 1998.
[18] Zhang D X, Zhang H, Zhao Y, et al. Applied Spectroscopy Reviews, 2022, 57(2): 89.
[19] Brunnbauer L, Gajarska Z, Lohninger H, et al. TrAC-Trends in Analytical Chemistry, 2023, 159: 116859.
[20] ZHANG Ran-ran, YING Lu-na, ZHOU Wei-dong(张冉冉, 应璐娜, 周卫东). Chinese Journal of Quantum Electronics(量子电子学报), 2023, 40(3): 376.
[21] Zhang T L, Yan C H, Qi J, et al. Journal of Analytical Atomic Spectrometry, 2017, 32(10): 1960.
[22] Wei P F, Lu Z Z, Song J W. Reliability Engineering & System Safety, 2015, 142: 399.
[23] DING Yu, YANG Lin-yu, CHEN Jing, et al(丁宇, 杨淋玉, 陈靖, 等). Laser & Optoelectronics Progress(激光与光电子学进展), 2022, 59(13): 1314006.
[24] Tavares T R, Mouazen A M, Nunes L C, et al. Soil & Tillage Research, 2022, 216: 105250.
[25] Li M G, Ruan F Q, Li R R, et al. Microchemical Journal, 2022, 182: 107928.
[26] Luarte D, Myakalwar A K, Velásquez M, et al. Analytical Methods, 2021, 13(9): 1181.
[27] Zhao Y J, Huo X M. Wiley Interdisciplinary Reviews-Computational Statistics, 2023, 15(4). https://doi.org/10.1002/wics.1602.
[28] Tognoni E, Cristoforetti G. Optics and Laser Technology, 2016, 79: 164.
[29] Hahn D W, Omenetto N. Applied Spectroscopy, 2012, 66(4): 347.
[30] Guezenoc J, Gallet-Budynek A, Bousquet B. Spectrochimica Acta Part B: Atomic Spectroscopy, 2019, 160: 105688.
[31] Syvilay D, Wilkie-Chancellier N, Trichereau B, et al. Spectrochimica Acta Part B: Atomic Spectroscopy, 2015, 114: 38.
[32] Yoon S, Choi J, Moon S J, et al. Applied Sciences-Basel, 2021, 11(15)：7154.
[33] Andries E. Journal of Chemometrics, 2013, 27(3-4): 50.
[34] Wong T T. Pattern Recognition, 2015, 48(9): 2839.