潜在表示学习框架下的低冗余多视图聚类算法

基于潜在空间学习的多视图聚类研究得到了较大发展,但其通常忽略了原始数据中冗余信息的存在可能会带来不理想的聚类结果。为解决这个问题,提出一种潜在表示学习框架下的低冗余多视图聚类算法。基于k-means的方法,直接从各视图数据学习其低维表示,由于该低维表示的各个特征相互正交,学习到的低维表示通常含有较少的冗余信息。基于潜在空间的假设,各视图的低维表示可由同一个潜在表示投影得到。将两者结合,就能得到一个具有低冗余信息的统一的潜在表示。设计了一个优化算法来求解目标问题,在多个公开数据集上的实验表明了该算法的有效性。     

高维稳健主成分聚类方法及其应用研究

随着信息技术的高速发展,每条数据所包含的信息越来越丰富,使得数据不可避免地含有异常值,且随着维数的增加,异常值出现的可能性更大。传统的主成分聚类分析对异常值特別敏感,基于MCD估计的主成分聚类方法虽然对异常值具有防御作用,但是在高维数据下MCD估计的偏差过大,其稳健性显著降低,而且当维数大于观测值个数时MCD估计失效。为此本文提出了基于MRCD估计的稳健主成分聚类方法,数值模拟和实证分析表明,基于MRCD估计的主成分聚类分析的效果优于传统的主成分聚类分析和基于MCD估计的主成分聚类分析,尤其是在维数大于样本观测值的情况下,MRCD估计更为有效。     

Uniqueness and clustering properties of Gibbs states for classical and quantum unbounded spin systems

We consider quantum unbounded spin systems (lattice boson systems) in -dimensional lattice space Z. Under appropriate conditions on the interactions we prove that in a region of high temperatures the Gibbs state is unique, is translationally invariant, and has clustering properties. The main methods we use are the Wiener integral representation, the cluster expansions for zero boundary conditions and for general Gibbs state, and explicitly -dependent probability estimates. For one-dimensional systems we show the uniqueness of Gibbs states for any value of temperature by using the method of perturbed states. We also consider classical unbounded spin systems. We derive necessary estimates so that all of the results for the quantum systems hold for the classical systems by straightforward applications of the methods used in the quantum case.     

模糊聚类应用于MIS子系统划分的研究

本文介绍了模糊聚类应用于ＭＩＳ子系统划分的新方法,包括基于模糊等价关系的方法和基于模糊相似关系的“最大树”方法．     

采用HK模型的防空反导作战网络建模分析

采用HK网络模型对防空反导作战网络进行建模,对网络度分布、集群系数、平均距离的统计特性进行分析。将网络模型与典型树状结构的防空反导网络和基于BA无标度模型的网络进行比较,表明基于HK模型的防空反导作战网络具有较好的集群系数和无标度特性,同时表明改变TF连接概率进行防空反导网络设计具有更好的灵活性。     

采用标准系统方块图实现OTA—C滤波器的设计

提出了一种采用现代控制理论中的状态空间表达式的系统方块图直接实现了ＯＴＡ－Ｃ滤波器的方法,ＳＰＩＣＥ程序模拟结果验证了设计方法的正确。     

Thermally Induced Measurement Error by a Water-Cooled Enthalpy Probe

Errors in stagnation-pressure measurement, due to a large temperature gradient at the face of a water-cooled enthalpy probe, were experimentally measured and numerically simulated. Two probes were used to measure the stagnation-pressure in a dc plasma jet; a standard water-cooled enthalpy probe and an uncooled ceramic (Thoria) probe. There was a maximum difference of 10% between the two measurements, with the water-cooled probe measuring lower pressures. Numerical simulations of plasma flow around the probe showed that the magnitude of the error depends on the thickness of the thermal boundary layer. The measurement error causes a maximum of 3% error in velocity measurements, using the Bernoulli equation. This error is no worse than other measurement errors associated with water-cooled enthalpy probe meaurements.     

Cooperative Competition Clustering for Gene Selection

Clustering analysis of data from DNA microarray hybridization studies is an essential task for identifying biologically relevant groups of genes. Attribute cluster algorithm (ACA) has provided an attractive way to group and select meaningful genes. However, ACA needs much prior knowledge about the genes to set the number of clusters. In practical applications, if the number of clusters is misspecified, the performance of the ACA will deteriorate rapidly. We propose the Cooperative Competition Cluster Algorithm (CCCA) in this paper. In the algorithm, we assume that both cooperation and competition exist simultaneously between clusters in the process of clustering. By using this principle of Cooperative Competition, the number of clusters can be found in the process of clustering. Experimental results on a synthetic and gene expression data are demonstrated. The results show that CCCA can choose the number of clusters automatically and get excellent performance with respect to other competing methods.     

A high-pressure mass spectrometric and density functional theory investigation of the thermochemical properties and structure of protonated dimers and trimers of glycine

A new modification of pulsed-ionization high-pressure mass spectrometry (PHPMS) has been used to perform equilibrium thermochemical studies for relatively nonvolatile biomolecules such as amino acids. Binding enthalpy and entropy changes have been measured for proton-bound clusters of glycine, which are in good agreement with both theoretical (DFT) results of this work and a previous blackbody infrared dissociation experiment. Experimental data indicate that a number of conformers of the proton-bound dimer of glycine may coexist in the explored temperature range (360-460 K). Several new, conceptually different isomers (two of them zwitterionic) have been found by DFT calculations, one of which is 7 kJ mol(-1) lower in energy than the structure previously reported to be the energy minimum.     

A study for quality evaluation of Taxilli Herba from different hosts based on fingerprint-activity relationship modeling and multivariate statistical analysis

In this study, a fingerprint-activity relationship modeling between chemical fingerprints and antirheumatic activity was established, and multivariate statistical analysis was used to evaluate the quality of Taxilli Herba (TH) from different hosts. Characteristic fingerprints of 20 batches of TH samples were generated by high-performance liquid chromatography coupled with triple quadrupole-time of flight tandem mass spectrometry (HPLC-Triple TOF-MS/MS), and the similarity analysis was calculated based on thirteen common characteristic peaks by hierarchical clustering analysis (HCA). Subsequently, nine efficacy markers were discovered by combining fingerprints and antirheumatic activity through grey correlation analysis (GCA) and bivariate correlation analysis (BCA). Meanwhile, the content of 5 constituents in 9 markers was determined by high-performance liquid chromatography coupled with triple quadrupole-linear ion trap tandem mass spectrometry (HPLC-QTRAP-MS/MS). The comprehensive quality of TH was assessed using multivariate statistical analysis, including principal components analysis (PCA) and technique for order preference by similarity to ideal solution (TOPSIS). The results showed that a high dose of TH extract could markedly ameliorate arthritis damage compared to other doses, with flavonoids playing an important role in the antirheumatic activity. The comprehensive quality of samples from Morus alba L. (SS) was superior to those from Liquidambar formosana Hance (FXS). The present study will demonstrate the markers associated with efficacy, and provide an applicable strategy for more comprehensive quality control and evaluation of TH.