Determination of glucose in plasma by dry film-based Fourier transformed-infrared spectroscopy coupled with boosting support vector regression.

In the present study, a dry film-based Fourier transformed-infrared (FT-IR) spectroscopic technique, coupled with boosting support vector regression (BSVR), was employed for a blood glucose assay. Potassium thiocyanate (KSCN) was taken in the dry-film method as an internal standard to compensate for any film thickness variation. This technique circumvents interference from water absorption, and requires only 5 microl of a sample. Moving window partial least-squares regression (MWPLSR) was used for wavenumber interval selection before multivariate modeling. By using the BSVR modeling technique, glucose in plasma could be determined over a 0.4 - 20 mmol/l concentration range with satisfactory accuracy. The performance of the BSVR methodology was compared with that of conventional support vector regression (SVR) as well as partial-least squares (PLS). The results demonstrated that BSVR is an effective multivariate calibration tool, providing better performance than conventional PLS and SVR.     

便携式X射线荧光光谱法结合支持向量回归算法定量分析土壤中的砷含量

研究了基于统计学习理论的支持向量机（SVM）回归法在X射线荧光光谱定量分析中的应用。以39个农田土壤样品作为实验材料,以其中32个土壤样品作为校正集,选用SVM模型中Linear、Poly和RBF 3种核函数对As元素含量与荧光光谱数据进行回归建模。用3种不同模型对预测集中7个土壤样品的As元素含量进行预测分析,结果显示模型预测As元素含量与电感耦合等离子体发射光谱法测定的As元素含量之间的相关系数R2均大于0.99,相对分析误差RPD均大于3,表明所建立的SVM模型具有较好的使用价值。为了进一步考察SVM回归模型的预测效果,同应用较成熟的PLS回归模型的预测结果进行对比,结果显示SVM法的预测结果更好,表明SVM回归模型亦可用于便携式X射线荧光光谱法的定量预测分析。     

Nonlinear regression by visualization of the sum of residual space applied to the integrated copolymerization equation with errors in all variables. I. Introduction of the model,simulations and design of experiments

A new model for estimating reactivity ratios using the integrated copolymerization equation is presented. The model is a general nonlinear least squares method taking the error in both monomer conversion and monomer fraction into account by a relation between these two variables. Simulations show that the model is able to predict reactivity ratios successfully. Special attention is given to experimental design, i.e., at which initial monomer feed ratios the experiments should be performed in order to obtain reliable values for the reactivity ratios. © 1999 John Wiley & Sons, Inc. J Polym Sci A: Polym Chem 37: 3793–3803, 1999     

Linking pore diffusivity with macropore structure of zeolite adsorbents. Part I: three dimensional structural representation combining scanning electron microscopy with stochastic reconstruction methods

In the present study we have employed advanced experimental and computational methods to characterize the structure of a binderless zeolite adsorbent with improved mass transfer characteristics. Hence, we have used standard and hybrid simulated annealing (SA) methods to stochastically reconstruct in three dimensions the adsorbent structure, by matching low order correlations, namely porosity and two point correlation function. These correlations are measured on two dimensional images obtained by Back-Scattered SEM micrographs taken on different samples of the adsorbent. In the standard SA method we have started from a purely random structure, while in the hybrid method we have started from a uniformly consolidated random sphere pack made by the Lubachevsky–Stillinger algorithm. It is found that the hybrid method preserves, besides low order correlations, pore and mass chord length distribution functions, which contain information on higher order correlations, while the standard SA method matches only low order correlations. This is because in the hybrid process we have initiated the SA algorithm starting from a structure that contains structural information of the sintered zeolite powder, in the form of a consolidated random sphere pack with the porosity of the final structure. Evidently, diffusion studies will enable us to further evaluate the structures developed by each method, as will be explored in a subsequent study.     

Separation of immunoglobulin G by high-performance pseudo-bioaffinity chromatography with immobilized histidine. I. Preliminary report on the influence of the silica support and the coupling mode.

High-performance liquid affinity chromatography with immobilized histidine as a pseudo-biospecific ligand has been used for the fractionation of human immunoglobulin G (IgG). Histidine was immobilized onto silica in two different modes: directly onto silica after epoxy activation or using an intermediate amino derivatization of silica and then coupling histidine using water-soluble carbodiimide. The behaviours and capacities of the obtained affinity supports as well as the influence of pH, silica type, pore diameter and coupling mode have been studied. IgG was effectively separated from human plasma and high maximal binding capacities were obtained.     

Metal complexes with ligands–organic phosphites as polyolefin antioxidants. I. Kinetics and mechanism of polyolefin and model hydrocarbon stabilization by organic phosphites

The most effective mechanism of protective action of organic phosphites–linear termination of kinetic chains on the inhibitor—has been established and the relative contribution of the other phosphite reactions (with hydroperoxides, their chain “self-oxidation”) to the summary mechanism has been determined. The quantitative kinetic characteristics of efficiency and length of inhibitive action of different structure phosphites in the wide range of the phase state of the oxidized substrate (solid polymer, its melt, liquid phase) have been given. On this basis the regularities which allow the prediction of the structures of phosphites scavenging peroxy substrate radicals more effectively (effective inhibitors) have been formulated.     

Nonlinear regression by visualization of the sum of residual space applied to the integrated copolymerization equation with errors in all variables. II. Application to the system methyl methacrylate-α-methylene-γ-butyrolactone using on-line Raman spectroscopy

The copolymerization of α-methylene-γ-butyrolactone and methyl methacrylate in DMSO was studied by on-line Raman spectroscopy. Reactivity ratios for this system were estimated from the in situ conversion measurements. The estimates are in good agreement with estimates obtained from low-conversion experiments where the composition of the copolymer was analyzed by ¹H-NMR. In order to obtain reliable estimates from the Raman data in combination with the integrated copolymerization equation, at least two experiments starting from different initial monomer feed fractions should be conducted. © 1999 John Wiley & Sons, Inc. J Polym Sci A: Polym Chem 37: 3804–3816, 1999