光谱结合主成分分析和模糊聚类方法的样品聚类与识别

针对紫外光谱结合化学计量学方法快速测定渣油烃族组成模型适应性问题,对渣油光谱进行主成分分析,以主成分得分作为聚类的特征变量进行模糊聚类,建立了光谱结合主成分分析和模糊聚类方法的样品聚类与识别方法和识别,为光谱结合化学计量分析方法中构正模型的正确选择提供了依据。     

基于分层聚类的黑色签字笔笔迹拉曼光谱研究

提出一种结合分层聚类和判别分析对笔迹成分进行分类检验的方法。利用激光显微共聚焦拉曼光谱仪对收集的市面上常见的130支黑色签字笔笔迹样本进行检测。对测量数据进行Savitzky-Golay卷积平滑和Z-score标准化处理,利用组间连接法、组内连接法和离差平方和法三种分层聚类方法对数据进行分类,将三种聚类方法所得分类结果作为判别依据进行判别分析,检验聚类方法的正确率。结合聚类树状图与正确率,最终选择在分类数为4时原始分类结果正确率为100%、留一交叉验证分类结果正确率为98.5%的离差平方和法,提出了适用于黑色签字笔笔迹拉曼光谱数据的分层聚类方法和判别验证方法。     

应用系统聚类法对海洛因裂解谱图分类

采用系统聚类法对海洛因样品裂解谱图进行分类研究.为了能够对铂丝裂解法分析海洛因溶液的可行性进行正确评价,首先对系统聚类法中的相似性量度和聚类方法进行了选择,然后运用相关性分析法和系统聚类法对裂解谱图的重现性结果进行测定,得到了较好的评价结果;最后对10种海洛因样品的裂解谱图进行了系统聚类分析,得出了分类结果.     

差分拉曼光谱技术结合K-means聚类法对牙膏的快速分类

建立了差分拉曼光谱技术结合K-means聚类法对牙膏快速分类的方法。对37个牙膏样品编号,将其分别涂抹于载玻片上,晾干,使用差分拉曼光谱仪进行扫描。调用R语言软件中fpc、factoextra、cluster数据库中的na.omit和scale函数对37个牙膏样品的差分拉曼光谱数据进行标准化处理,利用手肘法和Gap Statistic算法优化聚类数。在最佳聚类数为4的条件下,通过K-means聚类法对牙膏样品进行分类,并使用层次聚类分析法进行验证。结果显示,37个牙膏样品被分为4类,并且两种方法的分类结果一致。     

基于聚类分析对易燃液体塑料容器盖的光谱检验

将光谱分析方法与聚类方法相结合,对38个模拟案件环境的纵火现场助燃剂塑料容器盖样本进行X射线荧光光谱分析,根据元素种类将38个样本分为6组,建立了一种无损快速检验纵火现场物证的光谱分析方法。使用最短距离聚类法,以欧氏距离作为度量区间表示样本间亲疏关系,进行系统聚类分析,同时使用K-Means快速聚类法与t检验验证聚类效果。结果表明,38个样本分为6组时效果最好。并将各类的样本观测点在Fisher判别空间中的分布和位置情况绘出,与使用元素种类进行分类结果一致,为确定火灾案件性质,侦破案件线索提供了理论依据。     

聚类肽高分子材料及其结构与性能的研究

聚类肽又称为氮取代聚甘氨酸(N-聚甘氨酸),是一类具有优良生物相容性以及生物活性的可降解高分子材料.由于酰胺键的活泼氢被取代,聚类肽主链结构中消除了聚肽固有的多重氢键相互作用,其主链柔性较好,聚合物性质主要由侧链基团的种类及其物理化学性质决定.基于这种链结构特征,可以通过设计不同的侧基结构,有效地调节聚类肽高分子的热力学性能、降解性能和自组装行为等物理化学性质.合成聚类肽的方法主要有2种——开环聚合和固相合成.本文主要介绍了聚类肽高分子的本体与溶液自组装行为,系统阐明了如何通过调控聚类肽高分子的侧链结构,研究链结构与自组装行为之间的相互关系,进一步构筑具有独特相分离行为以及自组装结构的新型生物高分子,同时探讨了这些材料在生物医用和能源等领域的潜在应用.     

聚类分析法用于西洋参样品分类研究

建立西洋参药材分类方法;分别采用电感耦合等离子体质谱（ICP—MS）法、高效液相色谱（HPLC）法以及蒽酮-硫酸分光光度法,对12个西洋参样品中的15种无机元素、7种人参皂苷和多糖进行测定,用SPSS聚类分析法对12个西洋参样品进行聚类分析;结果表明：以4种不同的聚类变量进行聚类分析,其聚类分组结果并不一致,将几种有效成分综合为聚类变量,对西洋参进行分类更为科学合理;聚类分析法是西洋参分类分析的有效方法,对西洋参品种与质量的鉴定有一定的理论意义和实用价值。     

聚吡咯的电子能带结构

本文用EHMO方法计算了类苯型,类等键长型及类醌型等聚吡咯的电子能带.分析了原子电荷分布及价带,导带与能隙的宽度,并由此探讨了聚吡咯可能的电导机理.     

聚吡咯的电子能带结构

本文用EHMO方法计算了类苯型、类等键长型及类醌型等聚吡咯的电子能带.分析了原子电荷分布及价带、导带与能隙的宽度,并由此探讨了聚吡咯可能的导电机理.     

基于样本优化选取的光谱重建方法研究

针对光谱反射率重建中已有样本选取方法的不足,提出了一种基于核模糊C聚类的样本优化选取方法。该方法综合考虑了光谱反射率空间的广泛性和色度空间的相似性,较大程度满足了光谱重建的精度。首先采用已有样本选取法在光谱反射率空间选取C个样本作为聚类初始点,再将原光谱转化到色度空间进行聚类,同时引入核函数将二维色度空间映射到三维特征空间,使得特征数据线性可分,从而达到更好的划分效果。实验结果表明,使用该方法选取训练样本进行光谱反射率重建能够进一步提高光谱重建精度,色度评价和光谱评价结果均好于已有方法。     

ELECT++: Faster conformational search method for docking flexible molecules using molecular similarity

We have developed a program, ELECT++ (Effective LEssening of Conformations by Template molecules in C++), to speed up the conformational search for small flexible molecules using the similar property principle. We apply this principle to molecular shape and, importantly, to molecular flexibility. After molecules in a database are clustered according to flexibility and shape (FCLUST++), additional reagents are generated to screen the conformational space of molecules in each cluster (TEMPLATE++). We call these representative reagents of each cluster template reagents. Template reagents and clustered reagents produce, after reaction, template molecules and clustered molecules, respectively (tREACT++). The conformations of a template molecule are searched in the context of a macromolecular target. Acceptable conformational choices are then applied to all molecules in its cluster, thus effectively biasing conformational space to speed up conformational searches (tSEARCH++). In our incremental search method, it is necessary to calculate the root-mean-square deviations (RMSD) matrix of distances between different conformations of the same molecule to reduce the number of conformations. Instead of calculating the RMSD matrix for all molecules in a cluster, the RMSD matrix of a template molecule is chosen as a reference and applied to all the molecules in its cluster. We demonstrate that FCLUST++ clusters the primary amine reagents from the Available Chemicals Directory (ACD) successfully. The program tSEARCH++ was applied to dihydrofolate reductase with virtual molecules generated by tREACT++ using clustered primary amine reagents. The conformational search by the program tSEARCH++ was about 4.8 times faster than by SEARCH++, with an acceptable range of errors. © 1998 John Wiley & Sons, Inc. J Comput Chem 19: 1834–1852, 1998     

Accelerated K-means clustering in metric spaces

The K-means method is a popular technique for clustering data into k-partitions. In the adaptive form of the algorithm, Lloyds method, an iterative procedure alternately assigns cluster membership based on a set of centroids and then redefines the centroids based on the computed cluster membership. The most time-consuming part of this algorithm is the determination of which points being clustered belong to which cluster center. This paper discusses the use of the vantage-point tree as a method of more quickly assigning cluster membership when the points being clustered belong to intrinsically low- and medium-dimensional metric spaces. Results will be discussed from simulated data sets and real-world data in the clustering of molecular databases based upon physicochemical properties. Comparisons will be made to a highly optimized brute-force implementation of Lloyd's method and to other pruning strategies.     

Recursive intermediate factorization and complete computational linearization of the coupled-cluster single,double, triple,and quadruple excitation equations

Summary The nonlinear CCSDTQ equations are written in a fully linearized form, via the introduction of computationally convenient intermediates. An efficient formulation of the coupled cluster method is proposed. Due to a recursive method for the calculation of intermediates, all computational steps involve the multiplication of an intermediate with aT vertex. This property makes it possible to express the CC equations exclusively in terms of matrix products which can be directly transformed into a highly vectorized program.This work has been supported by the U.S. Air Force Office of Scientific Research, Grant No. 90-0079     

铁皮石斛的裂解气相色谱指纹图谱及其系统聚类分析

采用裂解气相色谱/质谱法(Py-GC/MS)测定了10种不同产地的铁皮石斛并结合系统聚类分析法比较了这些铁皮石斛的指纹图谱,采用释放气体分析法考察了裂解温度对指纹图谱的影响。结果表明,0.4 mg样品在450 ℃下可瞬间裂解,10种样品的指纹图谱具有相似性,且重现性好;采用系统聚类分析能区别不同产地的样品。本法快速、简便、准确,不失为药材质量控制的良好方法。     

激光产生的钒硫簇离子的统计分布与结构探讨

以脉冲激光在高真空中溅射钒粉和硫粉的混合物，产生了一系列钒硫原子簇正负离子。根据对其实验记录的激光等离子体质谱进行分析，发现钒硫原子簇的化学键基本上是共价型的，对应于相同钒原子数目的正负离子的相对丰度分别符合于不同形式的对数正态分布。正离子的相对丰度随硫原子数的增加呈对数衰减，由此可以确定各种大小簇合物的簇骼与几何构型；负离子的相对丰度则基本上呈正常的正态分布，说明在簇骼上结合了不同数量硫原子的负离子之结构稳定性大致相近。在实验中还观察到一类硫原子含量较高的簇离子，其中的部分硫原子可能在本身成键形成环状的硫集团后，又与原来的簇骼结合，因而在硫原子数较多时它们的相对丰度符合另一条对数正态分布曲线。     

Inkjet printing of gold nanoparticles onto a biologically relevant matrix to create quantitative test materials for time-of-flight secondary ion mass spectrometry

This study describes the use of inkjet printing for the preparation of test materials containing gold nanoparticles (AuNPs) on a biologically relevant matrix and discusses the methods of using time-of-flight secondary ion mass spectrometry (ToF-SIMS) for their spatially resolved quantification. Evaluation of test materials containing AuNPs with nominal diameters of (30, 80, 100, and 150) nm deposited onto gelatin with loadings ranging from 34 fg up to 67 000 fg per spot suggests that ToF-SIMS has the sensitivity and the dynamic range to quantify NP deposits in a biological matrix at toxicologically relevant concentrations, although it was not capable of reliably determining the size of the AuNPs from the intensity data. Regardless, the ability to extract intensity data from individual regions of interest (ROIs) showed that spatially resolved quantification is possible, even when multiple features exist in a single image and in a single depth profile. The argon gas cluster source used for sputtering led to a matrix removal effect where the matrix surrounding the AuNPs became negligible, which may facilitate the preparation of quantitative test materials.     

Approaching bulk limit for three‐dimensional solids via the cyclic cluster approximation: Semiempirical INDO study

The cyclic cluster method has been examined for a number of solids using a recently developed computer code, Solid 98. Calculations are based on the quasirelativistic (QR) INDO/1 (intermediate neglect of differential overlap) method that is simple enough to allow for a saturation of the (cyclic) clusters. Convergence toward the bulk limit (INDO/1) charge density with respect to the size of the model cyclic cluster is shown for diamond, silicon, germanium, boron nitride, gallium phosphide, gallium arsenide, and gallium antimonide. Results show that, as soon as the initial cluster size reaches 5 to 6 nm, one can safely use the obtained density matrix as a good approximation to the bulk limit. ©1999 John Wiley & Sons, Inc. J Comput Chem 20: 253–261, 1999     

On the calculation of expectation values and transition matrix elements by coupled cluster method

Summary Finite order expressions are derived for expectation values and transition matrix elements within the framework of the coupled cluster method.     

Topological Space of the Chemical Elements and its Properties

We carried out a mathematical study of 72 chemical elements taking advantage of the chemotopological method. We selected 128 properties to define the elements (physico-chemical, geochemical and chemical properties). Then, we looked for correlated properties and we reduced the number of them to 90. In this way we defined each element as a 90-tuple. Afterwards, we applied principal component analysis and cluster analysis (4 similarity functions and 5 grouping methodologies). Then, we calculated a consensus tree for the 20 dendrograms generated by the CA. Afterwards, we extracted the similarity relationships from the consensus tree and built up a basis for a topology on the set of chemical elements. Finally, we calculated some topological properties (closures, derived sets, boundaries, interiors and exteriors) of several subsets of chemical elements. We found that alkali metals, alkaline earth metals and noble gases appear not related to the rest of the elements. Also, we found that the boundary of non-metals are the semimetals with a stair-shape on the periodic table