一种新型非线性偏最小二乘方法研究及应用--基于Chebyshev多项式改进的PLS方法

在实际生产过程中，过程变量之间往往存在大量相关关系，甚至非线性相关关系。过程变量间存在线性相关时，可采用偏最小二乘方法(Partial Least Squares，PLS)计算模型参数，但由于PLS方法采用线性关系来联系输入和输出因子，因而并不能有效地应用于非线性较强的过程。在这种情况下要对变量进行有效的压缩维数．需要采用非线性PLS方法。基于Chebyshev多项式改进的多元多项式PLS方法，是一种新的非线性PLS方法。该方法利用Chebyshev多项式的正交性质和递推性质，将过程输入变量正交化、线性化后，再应用PLS方法计算模型参数并还原，从而得到比较精确的模型。由于该方法综合考虑了输入变量的自相关和输入变量间的协相关关系。能够更有效地表达过程变量间的非线性关系，因此其对非线性过程的研究提供了新的思路。     

Mapping of continuous floristic gradients in grasslands using hyperspectral imagery

Transitions between plant species assemblages are often continuous with the form of the transition dependent on the ‘slope’ of environmental gradients and on the style of self-organization in vegetation. Image segmentation can present misleading or even erroneous results if applied to continuous spatial changes in vegetation. Even methods that allow for multiple-class memberships of pixels presuppose the existence of ideal types of species assemblages that constitute mixtures—an assumption that does not fit the case of continua where any section of a gradient is as ‘pure’ as any other section like in modulations of grassland species composition.Thus, we attempted to spatially model floristic gradients in Bavarian meadows by extrapolating axes of an unconstrained ordination of species data. The models were based on high-resolution hyperspectral airborne imagery. We further modelled the distribution of plant functional response types (Ellenberg indicator values) and the cover values of selected species. The models were made with partial least squares (PLS) regression analyses. The realistic utility of the regression models was evaluated by full leave-one-out cross-validation.The modelled floristic gradients showed a considerable agreement with ground-based observations of floristic gradients (R²=0.71 and 0.66 for the first two axes of ordination). Apart from mapping the most important continuous floristic differences, we mapped gradients in the appearance of plant functional response groups as represented by averaged Ellenberg indicator values for soil pH (R²=0.76), water supply (R²=0.66) and nutrient supply (R²=0.75), while models for the cover of single species were weak.Compared to many other vegetation attributes, plant species composition is difficult to detect with remote sensing techniques. This is partly caused by a lack of compatibility between methods of vegetation ecology and remote sensing. We believe that the present study has the potential to increase compatibility as neither spectral nor vegetation information gets lost by a classifying step.     

A framework for joint estimation of age,gender and ethnicity on a large database

Human age, gender and ethnicity are valuable demographic characteristics. They are also important soft biometric traits useful for human identification or verification. We present a framework that can estimate the three traits jointly. The joint estimation framework could deal with the mutual influence of age, gender, and ethnicity implicitly. Under this joint estimation framework, we explore different methods for simultaneous estimation of age, gender, and ethnicity. The canonical correlation analysis (CCA) based methods, and partial least squares (PLS) models are explored under our joint estimation framework. Both the linear and nonlinear methods are investigated to measure the performance. We also validate some extensions of these methods, such as the least squares formulations of the CCA methods. We found some consistent ranking of these methods under our joint estimation framework. More importantly, we found that the CCA based methods can derive an extremely low dimensionality in estimating age, gender and ethnicity. An analysis of this property is given based on the rank theory. The experiments are conducted on a very large database containing more than 55,000 face images.     

Near-infrared spectroscopy to assist authentication and labeling of Asiago d’allevo cheese

This study investigated the possibility of using near-infrared spectroscopy (NIRS) for the authentication of Asiago d’allevo, a protected designation of origin cheese from northern Italy. Latent variable models applied on spectral data were developed and used to estimate several chemical properties and to classify the available dataset according to the location and management of the cheesemaking factory (lowland and alpine), the ripening age (6, 12, 18 and 36 months), the altitude of milk production (low, medium, medium–high and high), and the period of the year of the cheese production (May, July and September). The variable importance in projection index was used to identify the most informative spectral regions for discrimination. Results showed that NIR spectra can be used both to accurately estimate several chemical properties and to classify the samples according to the different experimental conditions under investigation with the same discrimination capacity provided by traditional chemical analysis.     

Modeling structure–activity relationships of prodiginines with antimalarial activity using GA/MLR and OPS/PLS

In the present study, we performed a multivariate quantitative structure-activity relationship (QSAR) analysis of 52 prodiginines with antimalarial activity. Variable selection was based on the genetic algorithm (GA) and ordered predictor selection (OPS) approaches, and the models were built using the multiple linear regression (MLR) and partial least squares (PLS) regression methods. The leave-N-out crossvalidation and y-randomization tests showed that the models were robust and free from chance correlation. The mechanistic interpretation of the results was supported by earlier findings. In addition, the comparison of our models with those previously described indicated that the OPS/PLS-based model had a higher quality of external prediction. Thus, this study provides a comprehensive approach to the evaluation of the antimalarial activity of prodiginines, which may be used as a support tool in designing new therapeutic agents for malaria.     

Monitoring of the manufacturing process for ambroxol hydrochloride tablet using NIR-chemometric methods: compression effect on content uniformity model and relevant process parameters testing

This study aimed at using near-infrared (NIR) spectroscopy to monitor compaction pressure for simultaneously determining the tensile strength and content uniformity, as well as moisture and mean particle size of ambroxol hydrochloride tablets. The content uniformity, compression force and tensile strength of the laboratory samples were obtained by pressing a mixture of active principle and excipient components into tablets. To reduce the spectral baseline shift of the laboratory samples, the compaction pressure applied to the mixture was assessed by a variable pressure test. Production samples were added to the test and subjected to principal component analysis. The expanded partial least-squares (PLS) calibration model used to quantify the active content was more accurate than the model constructed from laboratory samples using the production tablets included in the calibration set. The model showed good predictability, with correlation coefficient (R) 0.9977. The validation and reliability of the content model were evaluated to determine trueness and reliability for the measurement of individual production tablets and the laboratory tablets with drug content ranging from 24 to 36?mg. The PLS calibration models for compression force and tensile strength were constructed using the same spectral set assuming both were highly related. These models yielded high R values (0.9955 and 0.9910). The R values of the moisture and mean particle size were 0.9994 and 0.9919, respectively. This study demonstrated that NIR spectroscopy combined with chemometric techniques can be successfully used to quantitatively monitor the tablet manufacturing process in the pharmaceutical industry.     

一种基于可预测偏最小二乘法的故障检测方法

将可预测元分析(Fore CA)与偏最小二乘法(PLS)结合用于故障检测,在选取合适的可预测元的基础上,运用偏最小二乘回归,进一步提高模型对系统的预测能力,克服了偏最小二乘回归方法无法反映系统动态时序特性的缺陷,并构造CUSUM统计量和SPE统计量以检测故障是否发生。最后通过TE模型上的仿真实验结果表明:Fore PLS方法能有效检测慢漂移等故障。