In this work, we developed an analytical method based on UV-visible spectroscopy to determine the concentration of biodiesel from African palm in blends of petrodiesel. Seventy-five samples with biodiesel concentratio...In this work, we developed an analytical method based on UV-visible spectroscopy to determine the concentration of biodiesel from African palm in blends of petrodiesel. Seventy-five samples with biodiesel concentrations between 0-100 wt% were prepared. The spectral fingerprints that were obtained from the analysis of the samples by UV-visible spectroscopy were used to build predictive model using PLS regression. The predictive ability of the models was evaluated through statistical parameters: the standard error of calibration (SEC), the standard error of validation (SEV), the correlation coefficient of calibration (r Cal) and validation (r Val), the ratio (SEC/SEV), the coefficient of determination R2, the paired data Student’s t-test, cross-validation and external validation. The results indicate that the PLS model predicts the concentration of biodiesel from African palm with high precision in mixtures with petrodiesel. The method developed in this study can be applied to determine the concentration of biodiesel African palm in mixtures of petrodiesel in a more rapid and economical way. Moreover, this method has less analytical errors and is more environmentally friendly than the conventional methods.展开更多
极限学习机理论(extreme learning machine,ELM)作为一种新的化学计量学方法,在近红外光谱定量分析中的应用研究,已引起学术界的高度重视。然而,由于光谱数据维数较高,建立ELM模型时需要大量的隐节点,导致隐含层输出矩阵维数高且存在高...极限学习机理论(extreme learning machine,ELM)作为一种新的化学计量学方法,在近红外光谱定量分析中的应用研究,已引起学术界的高度重视。然而,由于光谱数据维数较高,建立ELM模型时需要大量的隐节点,导致隐含层输出矩阵维数高且存在高度共线性,用现有的Moore-Penrose广义逆算法求取隐含层输出矩阵与待测性质间的回归模型往往会存在病态问题。基于ELM建立光谱波长变量与性质之间的回归模型,提出以ELM模型隐含层输出矩阵作为新的变量,采用作者最新提出的基于变量投影重要性的改进叠加PLS算法(stacked partial least squares regression algorithm based on variable importance in the projection,VIP-SPLS),建立新变量与待测性质间的回归模型。VIP-SPLS算法充分利用了每个隐节点的输出信息,能有效解决高维共线性问题,同时具有模型集成的优点,从而改进了ELM模型的性能。将提出的改进ELM算法(improved ELM,iELM)应用于标准近红外光谱数据集,结果表明iELM模型的精度相对于现有的PLS模型和ELM模型分别显著提升了29.06%和27.47%。展开更多
文摘In this work, we developed an analytical method based on UV-visible spectroscopy to determine the concentration of biodiesel from African palm in blends of petrodiesel. Seventy-five samples with biodiesel concentrations between 0-100 wt% were prepared. The spectral fingerprints that were obtained from the analysis of the samples by UV-visible spectroscopy were used to build predictive model using PLS regression. The predictive ability of the models was evaluated through statistical parameters: the standard error of calibration (SEC), the standard error of validation (SEV), the correlation coefficient of calibration (r Cal) and validation (r Val), the ratio (SEC/SEV), the coefficient of determination R2, the paired data Student’s t-test, cross-validation and external validation. The results indicate that the PLS model predicts the concentration of biodiesel from African palm with high precision in mixtures with petrodiesel. The method developed in this study can be applied to determine the concentration of biodiesel African palm in mixtures of petrodiesel in a more rapid and economical way. Moreover, this method has less analytical errors and is more environmentally friendly than the conventional methods.
文摘极限学习机理论(extreme learning machine,ELM)作为一种新的化学计量学方法,在近红外光谱定量分析中的应用研究,已引起学术界的高度重视。然而,由于光谱数据维数较高,建立ELM模型时需要大量的隐节点,导致隐含层输出矩阵维数高且存在高度共线性,用现有的Moore-Penrose广义逆算法求取隐含层输出矩阵与待测性质间的回归模型往往会存在病态问题。基于ELM建立光谱波长变量与性质之间的回归模型,提出以ELM模型隐含层输出矩阵作为新的变量,采用作者最新提出的基于变量投影重要性的改进叠加PLS算法(stacked partial least squares regression algorithm based on variable importance in the projection,VIP-SPLS),建立新变量与待测性质间的回归模型。VIP-SPLS算法充分利用了每个隐节点的输出信息,能有效解决高维共线性问题,同时具有模型集成的优点,从而改进了ELM模型的性能。将提出的改进ELM算法(improved ELM,iELM)应用于标准近红外光谱数据集,结果表明iELM模型的精度相对于现有的PLS模型和ELM模型分别显著提升了29.06%和27.47%。
