循环空间回归为喹喏酮N1位构效关系建模

拟从线性泛函的角度分析循环子空间回归（CSR）。CSR方法将从自变量参数矩阵和因变量向量中提取成分,循环地构造并扩张Krylov子窨,且以此作为源空间,运用最小二乘准则解最映射到因变量实空间的线性泛函。整个求解过程包容了最小二乘回归（LSR）、主成分回归（PCR）、偏最小二乘回归（PLS）以及其它中间的回归方法。然后以预报能力的强弱,从中确定最佳回归模型。本文应用SCR方法为喹喏酮N1位抗菌构效关系建模,效果良好。     

循环子空间回归的一种快速算法及其在定量构效关系中的应用

循环子空间回归（ＣＳＲ）通过改变解空间的维数,可以获取一系列的回归模型,其中包括最小二乘回归（ＬＳＲ）、主成分回归（ＰＣＲ）、偏最小二乘回归（ＰＬＳＲ）和许多中间回归,从中可挑选最优回归模型。本文将分析ＣＳＲ的原理,给出一种可行的快速的ＣＳＲ算法（ＲＣＳＲ）,以提高计算速率和精度,并将其成功地应用于苯乙酰胺类除草剂定量构效关系的建模。     

组合偏最小二乘回归方法在近红外光谱定量分析中的应用

针对近红外光谱数据局部效应显著,变量个数多,彼此间常存在严重的复共线性,并多与样品组分含量呈非线性关系,构建一种组合非线性偏最小二乘回归(E-S-QPLSR)方法。它采用无重复采样技术(subag-ging),从训练样本中生成若干子样,然后每个子样通过二次多项式偏最小二乘回归(QPLSR),建立其子模型,并实现对训练样本因变量的定量预测,再将它们交由线性PLS算法用于计算各子模型的组合权系数。将该法应用于80个玉米样品的水组分含量与其近红外光谱的定量关系建模,效果良好,显示出很强的学习能力,所建模型的预报性能也优于其它方法。     

拉曼光谱同时测定乙醇与葡萄糖的方法研究

以共焦拉曼光谱技术为基础,结合96孔板,对比利用一元线性回归法、多元线性回归法、主成分回归法和偏最小二乘法建立定量分析模型,探索同时定量检测乙醇和葡萄糖的快捷方法。对同一组乙醇和葡萄糖混合标准溶液进行定量检测,并用木薯淀粉发酵液进一步验证。结果表明,偏最小二乘法定量结果的平均标准误差(SE)为0.179,平均相对标准偏差(RSD)为0.029,结合二阶导数后分别降至0.106和0.021,说明偏最小二乘法具有良好的紧密度和稳定性;t检验表明,在置信度为95%的条件下,实测值与标准值间不存在显著性差异。研究表明基于拉曼光谱技术的偏最小二乘回归定量方法,能满足实验和生产对乙醇和葡萄糖检测精度的要求,可用于乙醇和葡萄糖相关成分的同时快速检测。     

互信息诱导子空间集成偏最小二乘在近红外光谱定量校正中的应用

在集成框架下,提出了一种联合自助采样和基于互信息变量选择的子空间回归集成偏最小二乘算法MISEPLS.此算法的核心是通过训练集自助采样和随后计算互信息的方式来引入成员模型的差异性.由于互信息量小于一个特定阈值的变量被淘汰,每个成员模型在原始变量的一个子空间得到训练.模型融合考虑了简单平均和加权平均两种方式.通过两个近红外光谱定量校正实验,与建立单模型的全谱偏最小二乘算法(PLS)和基于互信息变量选择的偏最小二乘算法(MIPLS)进行了比较.结果表明,在不增加模型复杂度的情况下,MISEPLS能建立起更精确、更稳健的校正模型.     

两种梅花香气成分的分析及QSRR研究

采用固相微萃取(SPME)-气相色谱质谱(GC-MS)联用技术分析两种梅花的香气成分,通过保留指数与质谱解析相结合,分别对化合物进行定性分析.采用偏最小二乘回归(PLS)及多元线性回归(MLR)方法分别建立定量结构-色谱保留关系(QSRR)预测模型,并对训练集及测试集中化合物的保留指数进行预测.该研究为建立有效的GC-MS定性方法提供了一定的依据.     

构建支持向量机-偏最小二乘法为药物构效关系建模

为研究药物构效关系积累样本数据的过程中,需为小样本建模。此时较易造成过拟合,影响模型的预测性能和稳定性。为此可用偏最小二乘(PLS)法从样本数据中成对地提取最优成分,消除自变量间的复共线性,并有效的降维,然后应用最小二乘支持向量机对成对成分进行非线性回归,并以基于误差修正的策略调整,使之更有效地表达自、因变量间的非线性关系。由此构建为EB-LSSVM-PLS算法,所建模型的预报精度高,稳定性良好。将其应用于新型黄烷酮类衍生物的QSAR建模,效果令人满意,其泛化性能优于其它方法。     

基于机器学习方法的丙型肝炎病毒聚合酶NS5B非核苷抑制剂的定量构效关系研究

对89 个苯并异噻唑和苯并噻嗪类丙型肝炎病毒(HCV) NS5B聚合酶非核苷抑制剂进行了定量构效关系(QSAR)研究. 采用遗传算法组合偏最小二乘(GA-PLS)和线性逐步回归分析(LSRA)两种特征选择方法选择最优描述符子集, 然后建立多元线性回归和偏最小二乘线性回归模型. 并首次尝试使用遗传算法耦合支持向量机方法(GA-SVM)对两种特征选择方法所选的描述符子集分别建立非线性支持向量机回归模型. 三种机器学习方法所建模型均得到比较满意的预测效果. 采用LSRA所选的6 个描述符建立的三个QSAR模型对于测试集的相关系数为0.958-0.962, GA-SVM法给出最好的预测精度(0.962). 采用GA-PLS所选的7个描述符建立的三个QSAR模型对于测试集的相关系数为0.918-0.960, 偏最小二乘回归模型的结果最好(0.960). 本工作提供了一种有效的方法来预测丙型肝炎病毒抑制剂的生物活性, 该方法也可以扩展到其他类似的定量构效关系研究领域.     

基于LVs-LS-SVM近红外光谱检测柴油运动粘度的方法研究

为了研究不同光谱输入方法对柴油运动粘度预测模型的影响,本文对预处理后的柴油光谱进行主成分分析(PCA)得到的前6个主成分(PCs)、建立偏最小二乘(PLS)回归曲线选取有效波长(EWs)和运用偏最小二乘回归(PLSR)建模得到的14个潜变量(LVs),将三者分别输入至最小二乘支持向量机(LS-SVM),对结果进行比较分析表明:LVs-LS-SVM建模得到的结果为柴油运动粘度(R_(Pre)~2=0.839,RMSEP=0.317,RPD=1.54),优于EWs-LS-SVM和PCs-LS-SVM模型,为柴油参数测定便携式仪器的开发奠定基础。     

偏最小二乘法—流动注射pH梯度技术用于同时测定铜和钴

以PAR作显色剂,用流动注射pH梯度技术测定多个不同pH下的吸光度,以偏最小二乘法建立校正模型并预测,对Cu~(2+)、Co~(2+)二元素进行了同时测定,其计算结果优于主成分回归及多元线性回归法。     

Regression by L1 regularization of smart contrasts and sums (ROSCAS) beats PLS and elastic net in latent variable model

This paper proposes a regression method, ROSCAS, which regularizes smart contrasts and sums of regression coefficients by an L₁ penalty. The contrasts and sums are based on the sample correlation matrix of the predictors and are suggested by a latent variable regression model. The contrasts express the idea that a priori correlated predictors should have similar coefficients. The method has excellent predictive performance in situations, where there are groups of predictors with each group representing an independent feature that influences the response. In particular, when the groups differ in size, ROSCAS can outperform LASSO, elastic net, partial least squares (PLS) and ridge regression by a factor of two or three in terms of mean squared error. In other simulation setups and on real data, ROSCAS performs competitively. Copyright © 2009 John Wiley & Sons, Ltd.     

Numerical solution of ordinary differential equations by Fluctuationlessness theorem

This paper presents a numerical method based on Fluctuationlessness Theorem for the solution of Ordinary Differential Equations over appropriately defined Hilbert Spaces. We focus on the linear differential equations in this work. The approximated solution is written in the form of an nth degree polynomial of the independent variable. The unknown coefficients are obtained by setting up a system of linear equations which satisfy the initial or boundary conditions and the differential equation at the grid points, which are constructed as the independent variable’s matrix representation restricted to an n dimensional subspace of the Hilbert Space. An error comparison of the numerical solution and the MacLaurin series with the analytical solution is performed. The results show that the numerical solution obtained here converges to the analytical solution without using too many mesh points.     

Subspace Regression Ensemble Method Based on Variable Clustering for Near-Infrared Spectroscopic Calibration

Abstract

An ensemble approach, based on the combination of multiple linear regressions in subspace and variable clustering and therefore named VCS-MLR, was proposed for near-infrared spectroscopy (NIRS) calibration. By an experiment involving the determination of five components in tobacco samples, it was shown that VCS-MLR improved the performance by 61.4, 23.3, 10.2, 20.5, and 18, respectively, with respect to partial least squares regression (PLSR). The results confirmed that VCS-MLR can result in a more accurate calibration model but without the increase of computational burden. Moreover, the superiority of VCS-MLR was highlighted for small sample problems.     

子空间比较法研究拓扑块的变量关系及变量选择

用子空间比较法研究由不同拓扑指数构成的块变量之间的关系,提出了一种变量选择的新方法.以530种烷烃的5大类拓扑指数为样本进行比较,得到了反映拓扑块变量间线性相关关系的夹角余弦值,以及这些子空间所对应的结构信息.结合子空间比较的结果选择变量,对上述烷烃的沸点值进行回归,R=0.9948,S=4.08,交互检验预测误差LOO=4.38     

Variable selection in multivariate calibration based on clustering of variable concept

Recently we have proposed a new variable selection algorithm, based on clustering of variable concept (CLoVA) in classification problem. With the same idea, this new concept has been applied to a regression problem and then the obtained results have been compared with conventional variable selection strategies for PLS. The basic idea behind the clustering of variable is that, the instrument channels are clustered into different clusters via clustering algorithms. Then, the spectral data of each cluster are subjected to PLS regression. Different real data sets (Cargill corn, Biscuit dough, ACE QSAR, Soy, and Tablet) have been used to evaluate the influence of the clustering of variables on the prediction performances of PLS. Almost in the all cases, the statistical parameter especially in prediction error shows the superiority of CLoVA-PLS respect to other variable selection strategies. Finally the synergy clustering of variable (sCLoVA-PLS), which is used the combination of cluster, has been proposed as an efficient and modification of CLoVA algorithm. The obtained statistical parameter indicates that variable clustering can split useful part from redundant ones, and then based on informative cluster; stable model can be reached.     

Simultaneous Spectrophotometric Determination of Sunset Yellow and Quinoline Yellow in a Single Step

In this work a new modified standard addition method for simultaneous spectrophotometric determination of Sunset yellow (SY) and Quinoline yellow (QY) is used. The generalized Net Analyte Signal Standard Addition Method (GNASSAM) was introduced for simultaneous standard addition of analytes. By applying this method the concentrations of analytes are determined in a single step. The applicability of new method was used for the simultaneous determination of strongly overlapped spectra of SY and QY in a real sample without using any separation step. GNASSAM was also used to calculate figures of merit. Moreover, good limits of detection (SY=0.13 and QY=0.16 mg L ^?1) and suitable selectivity and sensitivity were determined. HPLC method was applied for finding subspace in binary mixtures of real samples. The results of investigated method for real sample were compared with HPLC results. The obtained results were in a good agreement with those of HPLC method.