Multivariate optimization of a GC-MS method for determination of sixteen priority polycyclic aromatic hydrocarbons in environmental samples

This paper describes the development and optimization, by using multivariate analysis, of a GC-MS-SIM method for evaluation of the 16 polyaromatic hydrocarbons considered as priority pollutants in atmospheric particulate material by the US EPA. In order to assure an adequate separation in the shortest analysis time, a multivariate design was used to set the conditions of the oven temperature program. The optimization process was carried out using factorial fractional design and Box-Behnken design. The following factors were evaluated: initial temperature, temperature rate #1, intermediary temperature, temperature rate #2, and final temperature. The optimized conditions were set at: 70 degrees C (2 min) --> 200 degrees C (30 degrees C/min, 5 min) --> 300 degrees C (5 degrees C/min, 1.67 min). Moreover, we have also optimized the injector temperature as 310 degrees C and sampling time as 0.8 min. The total analysis time was 33 min. Validation of GC-MS-SIM yielded satisfactory results for repetitivity of the detector response and retention times, and linearity of calibration curves. LOD were established as 0.13-0.34 ng/mL (peak area) and 0.18-0.72 ng/mL (peak height). The method has been shown to be appropriate for the analysis of samples of atmospheric particulate material and/or other environmental matrices.     

两个多元t分布之间的Kullback Leibler距离

该文将文［1］的结果推广到多元t分布的情形.结果表明, 两个多元t分布之间的Kullback Leibler距离与分布中的刻度矩阵之比的最大特征根密切相关.作为应用, 文章建立了一种在估计未知正定阵时的熵损失函数.     

多元弱样条空间和最小确定集

根据研究多元弱样条函数的B—网方法,给出了某些多元弱样条函数空间的最小确定集的构造方法,并从而求出了它们的维数．本文还讨论了对偶基的局部支集性质。