浏览全部资源
扫码关注微信
1.西安石油大学 化学化工学院,陕西 西安 710065
2.西北大学 化学与材料科学学院,陕西 西安 710127
李华,博士,教授,研究方向:化学信息学及能源分析化学,E-mail:huali@nwu.edu.cn
收稿日期:2025-01-18,
修回日期:2025-02-24,
录用日期:2025-02-28,
纸质出版日期:2025-08-15
移动端阅览
向宇,李茂刚,闫春华,张天龙,李华.近红外光谱结合LightGBM的含油污泥多环芳烃含量快速定量分析方法研究[J].分析测试学报,2025,44(08):1602-1611.
XIANG Yu,LI Mao-gang,YAN Chun-hua,ZHANG Tian-long,LI Hua.Rapid Quantitative Analysis of PAHs in Oily Sludge by Near-infrared Spectroscopy Combined with LightGBM[J].Journal of Instrumental Analysis,2025,44(08):1602-1611.
向宇,李茂刚,闫春华,张天龙,李华.近红外光谱结合LightGBM的含油污泥多环芳烃含量快速定量分析方法研究[J].分析测试学报,2025,44(08):1602-1611. DOI: 10.12452/j.fxcsxb.25011842.
XIANG Yu,LI Mao-gang,YAN Chun-hua,ZHANG Tian-long,LI Hua.Rapid Quantitative Analysis of PAHs in Oily Sludge by Near-infrared Spectroscopy Combined with LightGBM[J].Journal of Instrumental Analysis,2025,44(08):1602-1611. DOI: 10.12452/j.fxcsxb.25011842.
该研究利用近红外(NIR)技术结合轻量级梯度提升(LightGBM)预测含油污泥中菲(Phe)和荧蒽(Flt)的含量。首先对模型参数进行优化,然后对样品近红外光谱数据进行预处理,并通过竞争性自适应重加权算法(CARS)、互信息( MI)、鲸鱼优化算法(WOA)对光谱特征变量进行筛选,利用最优输入变量构建模型,最后将LightGBM与偏最小二乘(PLS)、随机森林(RF)、支持向量机(SVM)模型进行对比。结果表明,对于菲,基于Nor-SG-WOA-LightGBM建立的模型最佳,预测决定系数(
R
2
p
)和预测均方根误差(RMSE
p
)分别为0.995 2和0.242 6 mg/g;对于荧蒽,基于SNV-SG-CARS-LightGBM建立的模型最佳,
R
2
p
和RMSE
p
分别为0.995 1和0.245 2 mg/g。该方法为含油污泥中多环芳烃(PAHs)的分析提供了一定的技术参考。
In this study,a novel approach combining near-infrared(NIR) technology and the light gradient boosting machine(LightGBM) algorithm was developed for the quantitative prediction of phenanthrene and fluoranthene concentrations in oily sludge. Firstly,systematic optimization of model parameters was conducted to enhance predictive performance. Subsequently,spectral preprocessing techniques were applied to the NIR spectral data of the samples. Furthermore,the variable selection methods of competiti
ve adaptive reweighted sampling(CARS),mutual information(MI),and whale optimization algorithm(WOA) were used to select the spectral features effectively. Finally,based on the optimally selected input variables,the LightGBM prediction models were established. The performance of the LightGBM model was evaluated through comparative analysis with partial least squares regression(PLS),random forest(RF),and support vector machine(SVM). The results demonstrated that for phe content prediction,the Nor-SG-WOA-LightGBM model achieved superior performance,with a coefficient of determination(
R
²
p
) of 0.995 2 and root mean square error(RMSE
p
) of 0.242 6 mg/g. For flt content prediction,the SNV-SG-CARS-LightGBM model showed optimal performance,achieving an
R
²
p
of 0.995 1 and RMSE
p
of 0.245 2 mg/g. This method provides a technical reference for the analysis of polycyclic aromatic hydrocarbons(PAHs) in oily sludge.
Kang D Y , Lin H , Niu D P , Tian G , Xu R , Zhang Z Y , Luo Y J , Zhang Z Z . China Environ. Sci. (康定宇,林海,牛东坡,田刚,徐瑞,张志勇,罗一菁,张忠智. 中国环境科学), 2023 , 43 ( 8 ): 4106 - 4120 .
Poster D L , Schantz M M , Sander L C , Wise S A . Anal. Bioanal. Chem. , 2006 , 386 : 859 - 881 .
Feng G R , Li H , Huang X L , Huang D D , Chen L W , Cai X Q , Huang S Y , Hou M L . J. Instrum. Anal. (冯桂蓉,李红,黄先亮,黄丹丹,陈力维,蔡小钦,黄思瑜,侯美玲. 分析测试学报), 2024 , 43 ( 12 ): 2031 - 2042 .
Ćirić S , Mitić V , Jovanović S , Ilić M , Nikolić J , Stojanović G , Stankov Jovanović V . Microchim. Acta , 2018 , 185 : 1 - 10 .
Bao S , Qian J K , Fu H , Qiu T , Lu Y F . J. Instrum. Anal. (鲍珊,钱柬坤,付慧,邱天,陆一夫. 分析测试学报), 2024 , 43 ( 2 ): 261 - 268 .
Kim D Y , Lee B E , Shin H S . Anal. Biochem. , 2021 , 617 : 114119 .
Zacharioudaki D E , Fitilis I , Kotti M . Moleculars , 2022 , 27 ( 15 ): 4801 .
Dong G , Li X , Yang R , Yang Y , Liu H , Wu N . Environ. Pollut. , 2021 , 269 : 116150 .
Sakudo A . Clin. Chim. Acta , 2016 , 455 : 181 - 188 .
Ahmadi A , Emami M , Daccache A , He L . J. Agron. , 2021 , 11 ( 3 ): 433 .
Porep J U , Kammerer D R , Carle R . Trends Food Sci. Technol. , 2015 , 46 ( 2 ): 211 - 230 .
Zhang Y , Zhao Y J , Sun R R . Chin. J. Appl. Chem. ( 张怡,赵亚娟,孙儒瑞 . 应用化学) , 2016 , 45 ( S2 ): 262 - 264 .
Douglas R K , Nawar S , Alamar M C , Coulon F , Mouazen A M . Eur. J. Soil Sci. , 2019 , 70 ( 1 ): 140 - 150 .
Li M G , Feng Y Z , Yu Y , Zhang T L , Yan C H , Tang H S , Sheng Q L , Li H . Spectrochim. Acta A , 2021 , 257 : 119771 .
Ma C F , Zhai L L , Ding J M , Liu Y L , Hu S F , Zhang T L , Tang H S , Li H . Spectrochim . Acta A , 2024 , 310 : 123953 .
Ke G , Meng Q , Finley T , Wang T , Chen W , Ma W , Liu T Y . Adv. Neural Inf. Process. Syst. , 2017 , 30 : 52 .
Mao X , Ren N , Dai P , Jin J , Wang B , Kang R , Li D . Comput. Electron. Agric. , 2024 , 219 : 108818 .
HJ 805 - 2016 . Soil and Sediment-Determination of Polycyclic Aromatic Hydrocarbon by Gas Chromatography-Mass Spectrometry Method. The Environmental Protection Industry Standards of the People’s Republic of China(土壤和沉积物 多环芳烃的测定 气相色谱-质谱法. 中国环境保护行业标准) .
Sun J Y , Liu W X , Luo Y M , Ding L L , Ma T T , Huang Y J . Soils (孙剑英,刘五星,骆永明,丁琳琳,马婷婷,黄玉娟. 土壤), 2013 , 45 ( 1 ): 111 - 115 .
Li H , Liang Y , Xu Q , Cao D . Anal. Chim. Acta , 2009 , 648 ( 1 ): 77 - 84 .
Li Y , Yang X . Spectrochim. Acta A , 2023 , 285 : 121924 .
Li H , Li M , Tang H , Li H , Zhang T , Yang X F . RSC Adv. , 2023 , 13 ( 14 ): 9353 - 9360 .
Kraskov A , Stögbauer H , Grassberger P . Phys. Rev. E , 2004 , 69 ( 6 ): 066138 .
Mirjalili S , Lewis A . Adv. Eng. Softw. , 2016 , 95 : 51 - 67 .
He B , Jia B , Zhao Y , Wang X , Wei M , Dietzel R . J . Agric. Water Manage. , 2022 , 267 : 107618 .
Bian X H , Zhang R L , Liu P , Xiang Y , Wang S Y , Tan X Y . Spectrochim. Acta A , 2023 , 284 : 121788 .
0
浏览量
23
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构