基于波长选择方法Modeling Power的黑木耳产地判别研究

应用近红外光谱技术结合波长选择方法(modeling power,MP)实现了黑木耳产地的快速准确判别。共收集4个产地240个黑木耳样本,通过光谱扫描,建立了最优的偏最小二乘(PLS)判别模型。同时应用MP选择对黑木耳产地判别的有效波长,作为输入变量,建立最小二乘-支持向量机(MP-LS-SVM)模型。比较了3种MP选择波长的阈值方法,分别为MP值大于0.95,0.90和(0.90+Peak),并建立了相应的MP-LS-SVM模型。以预测集样本的准确判别率作为模型评价标准,分别设定预测的残差绝对值标准0.1,0.2和0.5。预测结果表明,MP-LS-SVM(0.90+Peak)模型在残差标准为0.1,0.2和0.5时的判别效果均为最优,正确判别率分别为98.3%,100%和100%。说明ModelingPower是一种非常有效的波长选择方法,应用近红外光谱技术结合MP-LS-SVM进行黑木耳产地判别是可行的,并获得了满意的判别精度。     

Potentiometric Electronic Tongue for Taste Assessment of Ibuprofen Based Pharmaceuticals

The flavor aspect of pharmaceutical formulations is very important in terms of their acceptability for the patient. This work is aimed at using of a novel promising type of electronic tongue (ET), consisting of a potentiometric sensor array, for the discrimination and assessment the taste of commercial pharmaceuticals based on ibuprofen (Ibuflam 4 %). The sensor array was formed by six ion‐selective membranes (ISMs) based on both specific and non‐specific active components. Analysis of potentiometric signal outputs including statistical data processing utilizing the principal component analysis (PCA) method allowed evaluating the role of individual active components of the ISMs in the sense of taste control of the pharmaceutical preparation Ibuflam 4 % before and after flavoring. The sensing array can be used to identify differences in flavor within individual batches from the same producer, since ISMs provide a different response to all components present in the analyzed pharmaceutical. The ET contains ISMs which are able to recognize the bitter taste in the framework of the batches of ibuprofen based pharmaceuticals. To recognize bitter taste, it was essential to include ISMs, providing a response to a particular type of taste, namely, specific and non‐specific ISMs. The experimental findings show that taste‐sensing potentiometric sensor system can be a good alternative to the human panel and is able to detect the slight changes in taste.     

荧光光谱法结合分子对接研究人血清白蛋白 对齐墩果酸与熊果酸的异构体识别作用

五环三萜化合物齐墩果酸(OA)与熊果酸(UA)为同分异构体,具有相似的理化性质和稍有差异的药理活性。目前人们主要采用各种色谱、质谱类方法实现对OA与UA的异构体识别,未见使用荧光光谱法的报道。提出了一种使用荧光猝灭法实现对OA与UA异构体识别的方法。首先考察了两种常见的血清蛋白-牛血清白蛋白(BSA)与人血清白蛋白(HSA)同OA与UA的作用情况,结果表明OA与UA均可有效地猝灭BSA与HSA的荧光发射。对所得荧光猝灭数据计算可知OA,UA与BSA,HSA作用的双分子猝灭速率常数(K_q)均远大于生物大分子荧光猝灭所观察到的最大散射碰撞速率常数2.0×1010 L·(mol·s)-1,说明猝灭类型均为静态猝灭,即OA与UA均是通过与BSA及HSA形成稳定复合物方式实现荧光猝灭的。应用双对数方程对所得荧光猝灭数据计算可知OA,UA与BSA,HSA所形成的复合物中结合位点数在0.90～1.26之间,说明所形成的复合物为1∶1型。BSA与OA,UA所形成复合物的表观结合常数(K_A)为同一数量级,相差不大,但是HSA与OA,UA所形成复合物的K_A差别很大,HSA-UA复合物的KA比HSA-OA复合物高124.91倍,表明HSA-UA复合物的稳定性更强。同步荧光实验结果显示,OA与UA的加入对于HSA波长差(Δλ)为60 nm同步荧光光谱的影响大于Δλ为15 nm的同步荧光光谱,由此可以说明OA与UA在HSA上的结合位点可能位于Trp残基附近。分子对接模拟计算结果表明OA与UA均对接在HSA结构中一个疏水性空腔中,主客体之间存在强烈的氢键与疏水作用。OA同Arg218,His242,Pro447等残基间存在氢键作用,键长分别为2.95,2.97与3.17 Å,此外还与Lys195,Lys199,Trp214,Arg222,Leu238,Asp451和Tyr452等七个氨基酸残基间存在疏水作用。UA同Trp214,Arg218和Lys444等残基存在氢键作用,键长分别为3.01,2.88与2.65 Å,此外还与Leu198,Gln221,Arg222,Asn295,Val343,Pro447,Cys448,Asp451和Val455等9个氨基酸残基间存在疏水作用。由于UA同HSA作用位点数目多于OA,说明UA与HSA疏水性空腔的空间匹配程度更高。因此,认为HSA-UA与HSA-OA复合物间稳定性差异是HSA实现对OA与UA异构体识别的原因。     

三维同步荧光光谱法快速辩别葵花籽油品质

利用三维同步荧光光谱法获取不同氧化状态下的葵花籽油荧光光谱数据,同时采集葵花籽油品质指标。运用平行因子法对三维同步荧光光谱矩阵进行降维处理,通过iPLS(interval partial least squares),BiPLS(backward interval partial least squares)和SiPLS(synergy interval partial least squares)模式识别方法进行数学建模。结果表明：波长差Δλ=50 nm时,样品同步荧光光谱具有显著差异,筛选用于数学建模初始数值。不同模式识别方法建模结果显示,iPLS,BiPLS和SiPLS法所得校正集模型和预测集模型的相关系数分别为0.908 3,0.961 2,0.954 5和0.872 3,0.925 2,0.852 5,交互验证均方根误差分别为0.050 3,0.033 1,0.035 9和0.073 3,0.054 1,0.065 5,比较发现采用BiPLS法建模效果最好。该研究将为葵花籽油品质快速辨别提供理论基础和技术支持,为其他食用油脂的快速检测提供方法指导。     

Customer rebates and retailer incentives in the presence of competition and price discrimination

Promotions are important tools for matching supply and demand in many industries. In the United States automotive industry, promotions are frequently offered, which may be given directly to customers (rebates) or given to dealers (incentives) to stimulate demand. We analyze the performance of customer rebate and retailer incentive promotions under competition. We study a setting with two manufacturers making simultaneous pricing and promotion decisions, and with two price-discriminating retailers as Stackelberg followers making simultaneous order quantity decisions. In the benchmark case with no promotions, we characterize the equilibria in closed form. We find that retailer incentives can be used by manufacturers to simultaneously improve each of their profits but can potentially lead to lower retailer profits. When manufacturers use customer rebates, we show that a manufacturer is able to decrease the profit of her competitor while increasing her own profit, although she is also at risk for her competitor to use rebates in a similar fashion. Unlike the monopoly case where the manufacturers are always better off with retailer incentives, customer rebates can be more profitable under some cases in the presence of competition. Using numerical examples we generate insights on the manufacturers’ preference of promotions in different market settings.     

Isotope (δ13C, δ15N, δ2H) diet–tissue discrimination in African grey parrot Psittacus erithacus: implications for forensic studies

Diet–tissue isotopic relationships established under controlled conditions are informative for determining the dietary sources and geographic provenance of organisms. We analysed δ¹³C, δ¹⁵N, and non-exchangeable δ²H values of captive African grey parrot Psittacus erithacus feathers grown on a fixed mixed-diet and borehole water. Diet–feather Δ¹³C and Δ¹⁵N discrimination values were +3.8?±?0.3?‰ and +6.3?±?0.7?‰ respectively; significantly greater than expected. Non-exchangeable δ²H feather values (?62.4?±?6.4?‰) were more negative than water (?26.1?±?2.5?‰) offered during feather growth. There was no positive relationship between the δ¹³C and δ¹⁵N values of the samples along each feather with the associated samples of food offered, or the feather non-exchangeable hydrogen isotope values with δ²H values of water, emphasising the complex processes involved in carbohydrate, protein, and income water routing to feather growth. Understanding the isotopic relationship between diet and feathers may provide greater clarity in the use of stable isotopes in feathers as a tool in determining origins of captive and wild-caught African grey parrots, a species that is widespread in aviculture and faces significant threats to wild populations. We suggest that these isotopic results, determined even in controlled laboratory conditions, be used with caution.     

需求弹性、抗衡势力与价格歧视

构建完全信息动态博弈模型,考察下游市场需求弹性、零售商抗衡势力对中间产品价格歧视的影响问题。结果表明,低弹性市场中零售商抗衡势力会削弱价格歧视,且改善低弹性市场消费者福利,而高弹性市场中零售商抗衡势力会增强价格歧视,且损害低弹性市场消费者福利。而高弹性市场消费者福利的变化则依赖于零售商的抗衡势力强度,只有当抗衡势力较强时,高弹性市场消费者福利才能够得到改善。     

基于小波特征的小麦白粉病与条锈病的定量识别

小麦白粉病和条锈病是小麦常发病害中为害较重的两种病害,在我国小麦产区均有发生,但它们由不同病原引起,需要采取不同的防治措施。因此,快速、准确的获取小麦病害类型信息对于病害的防治具有重要的指导意义。遥感数据具有快速、准确的获取空间上连续信息的特点,提出一种基于实测冠层高光谱数据信息的小麦病害定量识别方法。通过对标准化光谱进行连续小波变换,分析350～1 300 nm范围内各波段及其连续小波特征与小麦白粉病和条锈病之间的相关性,以及在不同病害间的差异性,筛选出对不同病害敏感的光谱波段(SBs)和小波特征(WFs),然后采用Fisher判别分析法分别基于SBs,WFs以及结合SBs和WFs建立小麦白粉病、条锈病及正常小麦识别模型,分别采用未参与建模的55个地面调查数据和留一法进行验证。结果显示： (1)基于WFs模型的总体识别精度(分别为92.7%和90.4%)明显高于基于SBs模型的总体识别精度(分别为65.5%和61.5%);(2)SBs和WFs结合模型的总体识别精度(分别为94.6%和91.1%)略高于基于WFs模型的总体识别精度,在Fisher80-55模型中白粉病和正常样本的生产者精度提高了10%以上。(3)条锈病样本能在基于WFs和SBs & WFs的模型中准确判别出来,用户精度和生产者精度均达到100%。结果表明采用作物光谱信息能够准确的识别健康作物和不同类型的作物病害,为采用遥感影像进行大范围作物病害识别提供了理论基础,对于指导作物病害防治具有实际应用价值。     

基于血流图和Fisher线性鉴别分析的红外人脸识别方法

为了从生物特征和统计角度来提高识别的性能,提出了一种基于血流图和Fisher线性鉴别(FLD)相结合的人脸识别方法,该方法首先利用血流模型把红外温谱图转换成血流图,能够利用人体的生物特征增加样本之间的类间距,并减少样本之间类内距,然后从统计特性出发,对血流图进行能最大化类     

酒香型光谱分析和模式识别计算分析

白酒是一个复杂的混合物体系,它含有大量的微量成分,这些微量成分直接决定了白酒的品质、口感和香型。为实现对白酒香型的快速鉴别,可采集不同香型白酒的红外光谱图,并将其作为模式分类方法的输入模式,建立白酒香型鉴别模型。首次全面系统地介绍了白酒香型模式识别算法,这些算法包括统计分类器(线性判别函数、二次判别函数、正则判别分析、K近邻算法)、原型学习算法(学习矢量量化)、支持向量机和AdaBoost算法。实验结果表明,基于红外光谱的白酒香型检测模式识别算法达到了很高的分类准确率、识别率和拒绝率,显示出了很好的性能。