用PCA-LDA对多环芳烃的致癌性进行分类

本文利用主成分提取-线性判别分析(PCA-LDA)模型对多环芳烃(Polycyclic Aromatic Hydrocarbons,PAHs)的致癌性进行分类,与致癌性有关的多环芳烃的表面积、代谢活性区域中心碳原子的离域能、亲电活性区域中心碳原子的离域能以及分子脱毒区的总数四个参数作为模型的输入,用已知致癌性的67个样本作为训练集建立PCA-LDA模型,对10个预测集样本的致癌性进行预测,结果表明:致癌性按高(h)、低(l)、非(n)分类时预测准确率达100%。     

基于RF_AdaBoost模型的血液种属鉴别算法

针对人类和非人类血液种属鉴别对无损、 高效分析方法的需求, 结合随机森林(Random Forest)和AdaBoost(Adaptive Boosting Algorithm)算法, 提出了一种血液种属鉴别方法(RF_AdaBoost). 该方法将RF作为AdaBoost的弱分类器, 以达到提高模型鉴别准确度, 增强模型鲁棒性的目的. 采用RF、 支持向量机(SVM)、 极限学习机(ELM)、 核极限学习机(KELM)、 堆栈自编码网络(SAE)、 反向传播网络(BP)、 主成分分析-线性判别法(PCA-LDA)及偏最小二乘判别分析(PLS-DA)与RF_AdaBoost模型进行对比, 以不同规模血液拉曼光谱数据训练集进行鉴别实验评估其性能. 结果表明, 随着训练样本的增加, RF_AdaBoost鉴别准确度最高达100%, 预测标准偏差趋于0. 与其它模型相比, RF_AdaBoost具有较高的分类准确度及较强的稳定性, 为血液种属的鉴别工作提供了新方法.     

基于遗传算法结合神经网络的航空润滑油粘度近红外光谱分析

研究了基于遗传算法(GA)的波长选择方法结合反向传播神经网络(BP-ANN)建模用于在用航空润滑油-40℃运动粘度的近红外光谱分析。采集样品光谱经均值中心化和SavitzkyGolay平滑求导法预处理后,通过分段组合建模初选最优波段,利用遗传算法进一步筛选了对粘度预测敏感的波长点建模。该波长选择方法与相关系数法相比,所建模型预测准确度高。在建模采用的非线性BP-ANN法中,先通过主成分分析(PCA)分解光谱数据,将得分矩阵输入3层神经网络训练,通过参数优化建立最优模型。所建模型对8个在用油进行分析,各预测值与标准值的相对误差均低于2%,并且经t检验不存在显著性差异,模型预测能力较强,应用于在用润滑油质量的快速分析效果好,为油品在线监控提供了参考。     

基于BP-神经网络的航空煤油总酸值近红外光谱快速检测

针对航空煤油中总酸值量较小,在近红外定量分析时有用信息易被干扰的问题,采用误差反向传播神经网络(BP-ANN)建立航空煤油总酸值近红外光谱分析模型.根据模型校正集预测偏差最小原则,确定了隐含层神经元个数、学习效率等参数.用建立的网络模型预测了验证集样品总酸值,预测的相关系数R2为0.9778,预测标准偏差(RMSEP)...     

遗传算法结合神经网络用于傅里叶变换红外光谱法测定航空润滑油中水分

采用傅里叶变换红外光谱法测定了航空润滑油中的水分,通过遗传算法(GA)优化选取有效波数点,用误差反向传播神经网络(BP-ANN)进行水分预测计算。模型的预测相关系数为0.957,预测标准偏差为0.022。随机抽取某型航空润滑油样品进行预测并对预测结果进行配对t检验,结果表明:红外光谱定量分析结果与标准方法测定值没有显著性差异,模型可以用于该型在用航空润滑油水分含量现场快速检测。     

基于粒子群算法的最小二乘支持向量机在红花提取液近红外定量分析中的应用

提出一种基于粒子群算法的最小二乘支持向量机(PSO-LS-SVM)方法,用于建立红花提取过程关键质控指标的定量分析模型.近红外光谱数据经波段选择、预处理和主成分分析(降维)后,利用粒子群优化(PSO)算法对最小二乘支持向量机算法中的参数进行优化,然后使用最优参数建立固含量和羟基红花黄色素A(HSYA)浓度的定量校正模型.将校正结果与偏最小二乘法回归(PLSR)和BP神经网络(BP-ANN)比较,并将所建的3个模型用于红花提取过程未知样本的预测.结果表明,BP-ANN校正结果优于PSO-LS-SVM和PLSR,但是对验证集和未知样品集的预测能力较差,而PSO-LS-SVM和PLSR模型的校正、验证结果相近,相关系数均大于0.987,RMSEC和RMSEP值相近且小于0.074,RPD值均大于6.26,RSEP均小于5.70％.对于未知样品集,pSO-LS-SVM模型的RPD值大于8.06,RMSEP和RSEP值分别小于0.07％和5.84％,较BP-ANN和PLSR模型更低.本研究所建立的PSO-LS-SVM模型表现出较好的模型稳定性和预测精度,具有一定的实践意义和应用价值,可推广用于红花提取过程的近红外光谱定量分析.     

高效液相色谱法结合化学计量学分析不同产地姜黄的指纹图谱

本文采用高效液相色谱法(HPLC)对来自于四川、广东和广西的68个姜黄样品进行质量控制方面的研究。通过高效液相色谱-质谱(HPLC-MS)联用技术对姜黄样品中的部分姜黄素类化合物和倍半萜烯类化合物进行了鉴定。进而对所采集到的姜黄样品的HPLC指纹图谱进行主成分分析(PCA)。结果表明,姜黄样品按产地被很好的分成了三类,其所含化合物的含量与产地有关。通过建立偏最小二乘判别分析(PLSDA)、反传人工神经网络(BP-ANN)及最小二乘支持向量机(LS-SVM)这三种有监督模式识别模型对未知样品进行产地预报,结果表明,非线性模型BP-ANN和LS-SVM的预报结果优于线性模型PLS-DA的结果。本文提出的方法可用于姜黄或其他一些中药或食品的质量控制,且一般不需要对它们的组分做定量分析。     

PCA-LVQ法及其在RS-FTIR大气环境监测数据处理中的应用

将主成分分析(PCA)用于遥感傅里叶变换红外光谱(Remote Sensing Fourier Transform Infrared:RS-FTIR)的特征提取,结合学习矢量量化(LVQ)神经网络,实现了PCA-LVQ对大气中的8组分混合体系进行快速定性分析的建模方法。并与单纯的LVQ神经网络、反向传播人工神经网络(BP-ANN)得到的结果进行了比较。PCA-LVQ显示出较好的处理数据的能力,它不仅提高了运算速度,而且提高了模型的预测准确度,分类精度达到91.7%。PCA-LVQ的这一预测精度及运算速度,足以满足遥感傅里叶变换红外光谱对大气中有毒气体的实时、在线监测的需要。     

近红外光谱法快速分析罗红霉素中两种主要成分

针对大环内酯类抗生素罗红霉素制剂的近红外光谱分析,利用后向区间偏最小二乘算法(biPLS)对全光谱进行波长区间选择预处理,并将其压缩为主成分,再利用BP人工神经网络(BP-ANN)建立预测模型,对其中的主要成分罗红霉素和乳糖进行预测。通过这个方法建立的模型,可以有效地抵御因为噪声等偶然因素引起的瞬时扰动,并在有效减少建模所用的波长数和建模运算时间的同时,使所建模型能达到较高的预测精度。为近红外快速在线同时检测药物多组分含量提供了新的参考方法。     

粒子群最小二乘支持向量机结合偏最小二乘法用于芝麻油质量的鉴别

结合粒子群最小二乘支持向量机(PSO-LSSVM)与偏最小二乘法(PLS)提出一种基于气相色谱技术的新方法,对芝麻油进行真伪鉴别,并对掺伪品中掺假比例进行定量分析。采用主成分分析法(PCA)对857个样本的脂肪酸色谱数据进行分析,优选主成分作为最小二乘支持向量机(LSSVM)的输入向量。利用粒子群算法(PSO)优化LSSVM,构建芝麻油掺伪鉴别的两级分类模型,同时运用PLS建立掺伪芝麻油中掺伪油脂的定量校正模型,两级分类模型的准确率分别达到了100%和98.7%,定量分析模型的平均预测标准偏差(RMSEP)为3.91%。结果表明,本方法的鉴别准确性和模型泛化能力均优于经典的BP神经网络和支持向量机(SVM),可用于食用油脂加工和流通环节的质量控制,为食用油质量的准确鉴定提供了一条有效途径。     

Recognition of benzene, toluene and xylene using TGS array integrated with linear and non-linear classifier

Three volatile organic compounds (VOCs): benzene, toluene and xylene were measured with an array of six Taguchi gas sensors in the air with variable humidity content. The recognition of single compounds was performed, based on measurement results. The principal component analysis (PCA) pointed at humidity as the main classification factor in the measurement data set. The linear discriminant analysis (LDA) was applied to overcome this drawback and enforce classification with respect to benzene, toluene or xylene. It was shown that discriminant function analysis (DFA), which is an LDA method allowed for 100% success rate in test samples recognition of benzene. It did not allow for accurate recognition of test samples of toluene or xylene. Following, the non-linear classifier, radial basis function neural network (RBFNN) was applied. A specific configuration of input ‘s was found, which provided for successful recognition of each single compound: benzene, toluene or xylene in air with variable humidity content.     

One- and two-dimensional gas chromatography–mass spectrometry and high performance liquid chromatography–diode-array detector fingerprints of complex substances: A comparison of classification performance of similar,complex Rhizoma Curcumae samples with the aid of chemometrics

Many complex natural or synthetic products are analysed either by the GC–MS (gas chromatography–mass spectrometry) or HPLC–DAD (high performance liquid chromatography–diode-array detector) technique, each of which produces a one-dimensional fingerprint for a given sample. This may be used for classification of different batches of a product. GC–MS and HPLC–DAD analyses of complex, similar substances represented by the three common types of the TCM (traditional Chinese medicine), Rhizoma Curcumae were analysed in the form of one- and two-dimensional matrices firstly with the use of PCA (Principal component analysis), which showed a reasonable separation of the samples for each technique. However, the separation patterns were rather different for each analytical method, and PCA of the combined data matrix showed improved discrimination of the three types of object; close associations between the GC–MS and HPLC–DAD variables were observed. LDA (linear discriminant analysis), BP-ANN (back propagation-artificial neural networks) and LS-SVM (least squares-support vector machine) chemometrics methods were then applied to classify the training and prediction sets. For one-dimensional matrices, all training models indicated that several samples would be misclassified; the same was observed for each prediction set. However, by comparison, in the analysis of the combined matrix, all models gave 100% classification with the training set, and the LS-SVM calibration also produced a 100% result for prediction, with the BP-ANN calibration closely behind. This has important implications for comparing complex substances such as the TCMs because clearly the one-dimensional data matrices alone produce inferior results for training and prediction as compared to the combined data matrix models. Thus, product samples may be misclassified with the use of the one-dimensional data because of insufficient information.     

Multivariate statistical analysis of tobacco of different origin, grade and variety according to polyphenols and organic acids

The aim of this work was to determine the concentration of polyphenols, organic acids in tobacco of different areas, grades and varieties by ultra-performance liquid chromatography tandem mass spectrometry (UPLC/MS/MS) and to achieve statistical classification by principal component analysis (PCA) and linear discriminant analysis (LDA). The obtained results revealed that tobacco of different varieties can be correctly classified according to the contents of polyphenols or organic acid. The results of PCA showed that different grades and geographic regions cannot completely be discriminated using polyphenols or organic acid as independent variables. However, there were marked differences in special class from the same type or grade tobacco. At the same time, the results of LDA also showed that the samples were correctly classified at 100% for different varieties of tobacco, but only 55.3% and 60% for different grades and areas, respectively. These results demonstrated that the composition of polyphenols and organic acids can be used as the useful variables to characterize the type and the special class or grade of tobacco.     

Classification of Chinese Herbal Medicine by Laser-Induced Breakdown Spectroscopy with Principal Component Analysis and Artificial Neural Network

Chinese herbal medicine has attracted increasing attention because of the unique and significant efficacy in various diseases. In this paper, three types of Chinese herbal medicine, the roots of Angelica pubescens, Codonopsis pilosula, and Ligusticum wallichii with different places of origin or parts, are analyzed and identified using laser-induced breakdown spectroscopy (LIBS) combined with principal component analysis (PCA) and artificial neural network (ANN). The study of the roots of A. pubescens was performed. The score matrix is obtained by principal component analysis, and the backpropagation artificial neural network (BP-ANN) model is established to identify the origin of the medicine based on LIBS spectroscopy of the roots of A. pubescens with three places of origin. The results show that the average classification accuracy is 99.89%, which exhibits better prediction of classification than linear discriminant analysis or support vector machine learning methods. To verify the effectiveness of PCA combined with the BP-ANN model, this method is used to identify the origin of C. pilosula. Meanwhile, the root and stem of L. wallichii are analyzed by the same method to distinguish the medicinal materials accurately. The recognition rate of C. pilosula is 95.83%, and that of L. wallichii is 99.85%. The results present that LIBS combined with PCA and BP-ANN is a useful tool for identification of Chinese herbal medicine and is expected to achieve automatic real-time, fast, and powerful measurements.     

表面解吸常压化学电离质谱法快速判别樟树化学型

采用表面解吸常压化学电离质谱(SDAPCI-MS)技术直接对5种化学型的樟树叶粉末片剂进行分析,获得其化学指纹谱图信息.采用主成分分析(PCA)、聚类分析(CA)和反向传输人工神经网络(BP-ANN)对谱图信息进行分析,获得各化学型樟树叶粉末片剂的特征质谱信息,进而对不同化学型样品进行判别.结果表明,在正离子模式下,SDAPCI-MS能快速获取樟树的化学指纹谱图;PCA分析中的PC1,PC2和PC3贡献率分别为79.9%,12.9%和4.2%,共计97.0%.SDAPCI-MS结合CA和BP-ANN测试样本准确率均为100%,能够快速、有效地判别出樟树化学型.     

偏最小二乘法在红外光谱识别茶叶中的应用

采用漫反射傅立叶变换红外光谱(FTIR)法结合主成分分析(PCA)、偏最小二乘法(PLS)、簇类的独立软模式(SIMCA)识别法对十三种茶叶进行了分类判别研究。研究结果表明,通过多元散射校正(MSC)对原始光谱进行预处理,可以提高模式识别技术的分类判别效果。在此基础上,选取1 900~900 cm-1波长范围内的茶叶红外光谱建立识别模型,三种方法都得到了满意的分类判别效果。在对检验集中全部130个样本的判别中,PCA仅有两类样本无法判别,SIMCA的识别率和拒绝率都在90%以上,而PLS的识别效果最佳,全部样本都得到了正确的归类。这一研究结果表明傅立叶变换红外光谱法与化学计量学方法相结合可以实现茶叶品种的快速鉴别,这为茶叶的客观评审提供了一种新思路。     

Application of multivariate methods to scalp hair metal data to distinguish between drug-free subjects and drug abusers

Human scalp hair samples of drug-free subjects and drug abusers (heroin and cocaine-heroin abusers) were analysed for trace metals by flame atomic absorption spectrometry (FAAS), flame atomic emission spectrometry (FAES) and electrothermal atomic absorption spectrometry (ETAAS). The classification of drug-free subjects and drug abuses groups with four multivariate methods using the metal contents in hair samples as discriminant variables has been discussed. Principal component analysis (PCA), cluster analysis (CA), linear discriminant analysis (LDA) and soft independent modelling of class analogy (SIMCA) allow distinguishing the two groups correctly. However, predictions by SIMCA are less satisfactory. Thirteen elements (Ag, Al, Ca, Cd, Cr, Cu, K, Mg, Mn, Na, Ni, Pb, and Zn) were determined by FAAS/FAES/ETAAS in 53 hair samples (16 samples of drug-free people and 37 samples of drug abusers). Human hair samples were prepared as aqueous slurries as sample pre-treatment and they were analysed using the slurry sampling technique. The half-range central value transformation was novelty used as data pre-treatment to homogenise the data. Grouping in the samples (drug-free people and drug abusers) were observed by using PCA and CA (squared Euclidean distance between objects and Ward method as clustering procedure). The application of LDA gave a correct recognition assignation percentage of 91.7 and 100.0% for the drug-free people and drug abusers, respectively, at a significance of 5%, while SIMCA offered recognition percentages of 83.3 and 91.3% for drug-free people and drug abusers, respectively, also at 5%. Finally, some studies were developed to classify heroin abusers and polidrug abusers (cocaine-heroin abusers) by the cited multivariate statistical methods. Recognition percentages of 90.9 and 100.0% were reached for heroin abusers and polidrug abusers groups, respectively, after LDA, while these percentages decreased to percentages lower than 90.0% when SIMCA was applied.