CAJ | 학술논문

表面增强拉曼散射（SERS）技术具有快速、指纹效应与极低的检测限等优点，被越来越多地应用到有害污染物、有毒物质、危险物质的检测与分析中。在SERS光谱的测量过程中，易受基底、仪器、宇宙射线与测量环境等因素影响，出现波动现象，对后续的分析与检测造成较大的干扰。基于农药福美双SERS光谱数据，尝试利用多种特征提取算法，如主成分分析（PCA）、离散余弦变换（DCT）、非负因式分解（NMF），对光谱的主分量进行提取，以减弱光谱数据波动对其后续的定量分析结果的影响。然后将提取后的分量分别结合线性回归算法———偏最小二乘法回归（PLSR），非线性回归算法———支持向量机回归（SVR）建立定量模型。最后，利用5-折交叉验证方法对比不同特征提取算法在不同类型的回归算法下的效果。通过实验验证可知，SVR对福美双溶液的分析精度要明显高于PLSR，这主要是由于SERS光谱强度与被分析物浓度之间为非线性关系。同时针对两种类型回归算法，特征提取算法都能明显地提升了分析结果，主要是由于其提取了源数据的主体信息，去除干扰信息。其中在线性回归中使用PCA效果最佳，在非线性拟合中使用NMF结果最佳，分析误差最好时可降低近3倍。最优回归模型（NMF+SVR）的交叉验证均方误差（RMSECV）为0．0455μmol·L－1（10－6 mol·L－1），达到国家对福美双的检测标准，为农药快速检测提供一种新的方法。
표면증강랍만산사（SERS）기술구유쾌속、지문효응여겁저적검측한등우점，피월래월다지응용도유해오염물、유독물질、위험물질적검측여분석중。재SERS광보적측량과정중，역수기저、의기、우주사선여측량배경등인소영향，출현파동현상，대후속적분석여검측조성교대적간우。기우농약복미쌍SERS광보수거，상시이용다충특정제취산법，여주성분분석（PCA）、리산여현변환（DCT）、비부인식분해（NMF），대광보적주분량진행제취，이감약광보수거파동대기후속적정량분석결과적영향。연후장제취후적분량분별결합선성회귀산법———편최소이승법회귀（PLSR），비선성회귀산법———지지향량궤회귀（SVR）건립정량모형。최후，이용5-절교차험증방법대비불동특정제취산법재불동류형적회귀산법하적효과。통과실험험증가지，SVR대복미쌍용액적분석정도요명현고우PLSR，저주요시유우SERS광보강도여피분석물농도지간위비선성관계。동시침대량충류형회귀산법，특정제취산법도능명현지제승료분석결과，주요시유우기제취료원수거적주체신식，거제간우신식。기중재선성회귀중사용PCA효과최가，재비선성의합중사용NMF결과최가，분석오차최호시가강저근3배。최우회귀모형（NMF+SVR）적교차험증균방오차（RMSECV）위0．0455μmol·L－1（10－6 mol·L－1），체도국가대복미쌍적검측표준，위농약쾌속검측제공일충신적방법。
Three feature extraction algorithms,such as the principal component analysis (PCA),the discrete cosine transform (DCT)and the non-negative factorization (NMF),were used to extract the main information of the spectral data in order to weaken the influence of the spectral fluctuation on the subsequent quantitative analysis results based on the SERS spectra of the pesticide thiram.Then the extracted components were respectively combined with the linear regression algorithm—the partial least square regression (PLSR)and the non-linear regression algorithm—the support vector machine regression (SVR)to devel-op the quantitative analysis models.Finally,the effect of the different feature extraction algorithms on the different kinds of the regression algorithms was evaluated by using 5-fold cross-validation method.The experiments demonstrate that the analysis re-sults of SVR are better than PLSR for the non-linear relationship between the intensity of the SERS spectrum and the concentra-tion of the analyte.Further,the feature extraction algorithms can significantly improve the analysis results regardless of the re-gression algorithms which mainly due to extracting the main information of the source spectral data and eliminating the fluctua-tion.Additionally,PCA performs best on the linear regression model and NMF is best on the non-linear model,and the predic-tive error can be reduced nearly three times in the best case.The root mean square error of cross-validation of the best regression model (NMF+SVR)is 0. 045 5μmol·L-1 (10-6 mol·L-1 ),and it attains the national detection limit of thiram,so the meth-od in this study provides a novel method for the fast detection of thiram.In conclusion,the study provides the experimental ref-erences the selecting the feature extraction algorithms on the analysis of the SERS spectrum,and some common findings of fea-ture extraction can also help processing of other kinds of spectroscopy.