葡萄糖水溶液FTIR/ATR光谱分析的吸收峰组合模型

利用傅立叶变换红外光谱（FTIR）结合衰减全反射（ATR）技术建立葡萄糖水溶液的分析模型.基于葡萄糖水溶液与去离子水的差谱,找到5个吸收峰：1150,1103,1078,1034,991（cm-1）.建立吸收峰组合多元线性回归（MLR）模型,最优组合为1103,1034,991（cm-1）,对应预测均方根偏差（RMSEP）、预测相关系数（RP）分别为0.933 mmol/L,0.989,大幅度优于全谱（4500-600 cm-1）偏最小二乘（PLS）模型的预测效果,并且大大降低了模型复杂性.为了得到稳定可靠的结果,所有模型都是基于50个定标集和预测集划分的平均效果得到的.     

Source Apportionment and Risk Assessment of Polycyclic Aromatic Hydrocarbons (PAHs) in Surface Sediments from Upper Reach of Huaihe River,China

Fifteen priority PAHs in 16 sediments from upper reach of Huaihe River collected in 2007 were measured. Source apportionment derived from PCA/MLR indicated that the highest contribution to ΣPAHs was from biomass burning (34.1%), followed by coal combustion (29.7%), vehicular emission (25.8%), and coke oven origin (10.4%). Risk assessment for each identified source was quantitatively calculated by combining the benzo[a]pyrene equivalent (BaPE) with estimated source contribution. The results showed that coal combustion posed the highest toxic risk, with BaPE value of 13.1 ng g^?1 dw, and the BaPE values for biomass burning, vehicular emission and coke oven were 5.6 ng g^?1 dw, 7.8 ng g^?1 dw, and 4.1 ng g^?1 dw, respectively. The distributions of contribution for total PAHs burden and BaPE of each identified source showed similar pattern among different sampling sites.     

基于改进的多元离群检测方法的风机齿轮箱早期故障诊断

针对风电机组运行工况波动性以及机组早期故障特征不易提取的特点,提出一种基于改进的多元离群监测方法来实现风机齿轮箱故障的早期诊断。运用阶比重采样方法对原始振动信号进行预处理,并对处理结果进行量纲一因子分析;通过马氏距离建立风电齿轮箱的早期故障识别模型;利用多元线性回归改进多元离群检测算法进行实际数据的分析计算。结果表明,该方法较原始方法能够更早地察觉出风电齿轮箱早期故障。     

无偏估计拆分算法在数控加工中心主轴热误差建模中的应用

针对多元线性回归无偏估计算法在处理具有多重共线性的机床热误差数据建模中出现的模型参数估计失真问题,提出了一种用于处理共线性数据的无偏估计拆分算法。该算法将建模过程分成多个步骤完成,每步只对一个自变量进行回归,从而达到弱化自变量共线性的目的。以Leaderway-V450型数控加工中心为实验对象,根据在不同季度内测量的多批次空转实验数据,将无偏估计拆分算法与传统多元线性回归的模型精度和稳健性进行了验证。研究结果显示,无偏估计拆分模型的预测精度和稳健性远优于经典多元线性回归模型,尤其对于跨季度数据预测,该算法优势更大。     

A novel hybrid classification model of artificial neural networks and multiple linear regression models

The classification problem of assigning several observations into different disjoint groups plays an important role in business decision making and many other areas. Developing more accurate and widely applicable classification models has significant implications in these areas. It is the reason that despite of the numerous classification models available, the research for improving the effectiveness of these models has never stopped. Combining several models or using hybrid models has become a common practice in order to overcome the deficiencies of single models and can be an effective way of improving upon their predictive performance, especially when the models in combination are quite different. In this paper, a novel hybridization of artificial neural networks (ANNs) is proposed using multiple linear regression models in order to yield more general and more accurate model than traditional artificial neural networks for solving classification problems. Empirical results indicate that the proposed hybrid model exhibits effectively improved classification accuracy in comparison with traditional artificial neural networks and also some other classification models such as linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), K-nearest neighbor (KNN), and support vector machines (SVMs) using benchmark and real-world application data sets. These data sets vary in the number of classes (two versus multiple) and the source of the data (synthetic versus real-world). Therefore, it can be applied as an appropriate alternate approach for solving classification problems, specifically when higher forecasting accuracy is needed.     

5′-甲硫腺苷核苷酶抑制剂的QSAR研究

通过构建33个甲硫腺苷(MTA)类似物的化学结构模型,并使用Molconn-Z统计的电子拓扑指数、形状指数和分子连接性指数作为受试化合物分子结构参数,结合多元线性回归(MLR)方法对MTA类似物的抑制活性进行了QSAR研究.所得模型的R和R2adj分别为0.862和0.711,表明该模型具有较好的预测和反映真实情况的能力.模型表明,MTA类似物的抑制活性与其所含硫和氮原子的电子拓扑指数以及分子形状指数相关,这为预测其他类似物的活性和设计新的抑制剂提供了理论依据.     

Computational screening and design of traditional Chinese medicine (TCM) to block phosphodiesterase-5

The traditional Chinese medicines (TCM), Epimedium sagittatum (ESs), Cnidium monnieri (CMs), and Semen cuscutae (SCs), were used for treating erectile dysfunction since the ancient Han dynasty (202 BC–AD 220). Phosphodiesterase-5 (PDE-5) is deemed the target protein for inhibition to treat erectile dysfunction. In this study, a reliable multiple linear regression (MLR) model (r value = 0.8484) was used to predict the activities of new candidates which were designed from ES, CM, and SC. From docking and pharmacophore analysis, the potent candidates among ES, CM, and SC were screened. SC01, SC03, and ES03b were predicted to have high potencies based on MLR analysis and high docking scores. Additionally, from our analysis, we make the follow conclusion (1) Hydrophobic compounds tend to be more potent PDE-5 inhibitors; (2) Because of the big binding site, inhibitors with molecular weights over 500 remain potent; (3) From the pharmacophore analysis, the features of hydrogen bond acceptors are the basis for designing novel inhibitors of PDE-5 and (4) According to MLR analysis, the number of ring groups could be up to 6, but the number of aromatic rings was limited to 4 to be potent. This study offers an alternative way to screen PDE-5 inhibitors from TCM and provides a scientific basis for confirming pharmacological actions of TCM.     

Rapid assessment of neoformed compounds in nuts and sesame seeds by front-face fluorescence

The potential of a rapid spectral method, based on front-face fluorescence, to monitor lipid neoformed compounds (NFC) during processing of nuts and sesame seeds was investigated. Fluorescence fingerprints were obtained from front-face fluorescence acquisition directly on crushed nuts and sesame seed samples obtained at different stages of processing. Fluorescence was very sensitive to physicochemical changes induced by the heat process, namely roasting. Parallel factor (PARAFAC) analysis of the fluorescence landscapes revealed four main fluorescence profiles in the nuts, and five in the sesame seeds. These were associated with peptidic tryptophan, tocopherols and process derived products. Various regression models between fluorescence spectra and NFC appearing during the process, carboxymethyllysine (CML) and trans fatty acids (tFA) showed good correlations (R > 0.89) and satisfactory prediction errors (RMSECV < 1.67). When applied to indicators of lipid peroxidation, good regression models were also obtained allowing prediction of the pAV (p-anisidine value) and TBARs (thiobarbituric acid reactive substances): R = 0.73 and 0.96 in nuts and sesame seeds, respectively, with prediction errors lower than 0.78. This study demonstrates the interest of front-face fluorescence as a promising tool for quality control of nuts and seeds roasting.