Nonlinear discriminant clustering based on spectral regularization

Owing to sparseness, directly clustering high-dimensional data is still a challenge problem. Therefore, obtaining their low-dimensional compact representation by dimensional reduction is an effective method for clustering high-dimensional data. Most of existing dimensionality reduction methods, however, are developed originally for classification (such as Linear Discriminant Analysis) or recovering the geometric structure (known as manifold) of high-dimensional data (such as Locally Linear Embedding) rather than clustering purpose. Hence, a novel nonlinear discriminant clustering by dimensional reduction based on spectral regularization is proposed. The contributions of the proposed method are two folds: (1) it can obtain nonlinear low-dimensional representation that can recover the intrinsic manifold structure as well as enhance the cluster structure of the original high-dimensional data; (2) the clustering results can also be obtained in the dimensionality reduction procedure. Firstly, the desired low-dimensional coordinates are represented as linear combinations of predefined smooth vectors with respect to the data manifold, which are characterized by a weighted graph. Then, the optimal combination coefficients and the optimal cluster assignment matrix are computed by maximizing the ratio between the between-cluster scatter and the total scatter simultaneously as well as preserving the smoothness of the cluster assignment matrix with respect to the data manifold. Finally, the optimization problem is solved in an iterative procedure, which is proved to be convergent. Experiments on UCI data sets and real world data sets demonstrated the effectiveness of the proposed method for both clustering and visualization high-dimensional data set.     

面向大规模数据集的并行化Top-k Skyline查询算法

随着数据规模的日益庞大,在大规模数据集中帮助用户定位出数据量可控的代表性信息显得越发重要。虽然Top-k Skyline查询能够找到数据集中前k个最具代表性的信息,在获取代表性信息的同时又控制了结果规模,满足了上述要求,但是现有的Top-k Skyline查询在面对大规模数据集时效率较低,并不适用于大规模数据集。为了解决这个问题,将Top-k Skyline查询与并行化处理相结合,提出了一种面向大规模数据集的并行化Top-k Skyline查询算法PTKS（parallel Top-k Skyline）,通过充分利用分布式资源,将原有查询进行有效的并行化处理,同时设计了基于用户偏好的用于缩减结果数据量的筛选规则,满足用户需求。在真实数据集上进行了相关实验,并与现有方法进行了对比,结果表明PTKS在大规模数据集上的查询效率更具有优势,能很好地适用于大规模数据集。     

Experimental Study on Slope Runoff,Erosion and Sediment under Different Vegetation Types

The relationships between precipitation, vegetation and erosion are important yet unresolved issues in the field of earth surface processes. Vegetation plays an important role in controlling soil erosion. Through field simulated rainfall experiments, we analyzed the characteristics, regulation of, and correlation among the slope rainfall-infiltration-runoff, erosion and sediment under different vegetation types. The results showed that the forest effectively improved soil structure, had stronger runoff and sediment regulation and was influenced less by rainfall intensity than those under other vegetative conditions. In addition, the efficiency and pattern of the regulation of runoff and sediment varied with vegetation types as did the mechanism of action. The soil and water conservation function of forest was water storage and sediment reduction by plant root systems to reduce erosion power, increase infiltration, decrease runoff and reduce flow speed. The function of grassland was direct sediment interception based on surface vegetation canopy for runoff and sediment regulation. The root contribution to runoff and sediment reduction was relatively greater than the shoot contribution under forest conditions, whereas, the effect of shoots and roots on soil loss was almost equivalent under grassland conditions. The different spatial structures of vegetation affected runoff and sediment regulation in different ways, and plant root systems were crucial for soil and water conservation. The cumulative sediment yield of the slopes increased as a statistically significant power function of cumulative runoff. The coefficient and curve shape of function were dependent on vegetation type, soil properties, rainfall intensity and surface roughness. The process of slope runoff and sediment was divided into development, active and stable stages. These stages correlated with each other to constitute a complete rainfall-runoff and erosion-sediment process, which exhibited their own features at each stage. This study furthers understanding of the relationships between vegetation, soil erosion and precipitation.     

多特征证据理论融合的视觉词典构建

现有的视觉词典构建方法一般是将多个特征构成整个向量并通过聚类形成视觉词典,这种方法在视觉聚类的过程中只考虑了特征的整体相似性而忽略了不同特征对构建视觉词典的影响。提出了一种基于Dempster-Shafer（D-S）证据理论的多特征融合的视觉词典构建方法。该方法应用证据理论融合不同特征的视觉相似性,构造出更加精确的视觉词典。在证据理论的基础上,使用两种特征实现了对视觉词典的再分,使得相似的特征更好地集中在一类中。与传统视觉词典构建方法相比,本文方法获得了更好的结果。应用以上视觉词典构造方法并将之应用于分类实验,在支持向量机与朴素贝叶斯分类器上取得的分类实验结果表明,应用本文方法构建的视觉词典能有效提高视觉词典的精确度,分类效果得到了很大的提高。     

基于Bootstrap区间估计的中药指纹图谱相似度阈值确定方法研究

利用指纹图谱相似度评价中药质量时,色谱指纹图谱样本的数量有限。小样本条件下指纹图谱相似度值的分布难以满足正态分布或近似正态分布,不适合直接利用传统的区间估计方法来确定相似度的阈值。本文利用Bootstrap方法扩增样本,建立了小样本条件下的指纹图谱相似度阈值确定方法,并通过3个具体实例说明指纹图谱相似度阈值确定的具体过程。首先通过Bootstrap方法扩增样本,然后利用单侧区间估计方法确定相似度下限,得到相似度的阈值。以上3个实例说明了用该方法建立中药指纹图谱相似度阈值的合理性和可行性。此方法适用于由不同相似度计算方法所得相似度值。本文从理论、技术和应用的角度阐述了确定中药指纹图谱相似度阈值的方法,该方法将对指纹图谱技术的推广应用产生积极作用。     

数字隔离型RS-232转M-Bus接口电路的设计

分析介绍了仪表总线(M-Bus)信号和传统的RS-232转M-Bus的电路实现方式,并就此分析了光电耦合隔离型电路和数字隔离型电路各自的优缺点,在此基础上提出了采用了ADI公司数字隔离芯片ADuM1201来实现数字隔离型RS-232转M-Bus的电路设计.这种新型电路简单易用、隔离性能更强、功耗更低并具有很好的可靠性.     

基于CVaR准则具有预算约束和损失约束的报童决策

为了探讨风险态度、现实约束对决策行为的影响, 建立了具有预算约束、损失约束、需求依赖于价格、基于条件风险价值(CVaR) 准则的价格-订单量决策模型, 并给出了模型最优解与约束条件有效阀值求解方法. 研究结果表明, 条件风险价值对损失约束比对预算约束的敏感程度高, 而不同约束的高敏感区域位置不同; 风险规避程度越高, CVaR 对约束值的弹性系数越低; 预算约束值的提高降低了边际贡献率, 风险规避程度越高, 降低幅度越大.

     

Dissipative control for linear systems by static output feedback

In this article, the problem of static output-feedback dissipative control is investigated for linear continuous-time system based on an augmented system approach. A necessary and sufficient condition for stability and strict (Q,S,R)-dissipativity of the closed-loop system is established in terms of a matrix inequality with free parametrisation matrix. An equivalent characterisation with some slack matrices for numerical solvability is then proposed. Based on this, a necessary and sufficient condition for the existence of a desired controller is given, and a corresponding iterative algorithm is developed to solve the condition. The effectiveness of results developed in this article is demonstrated by some numerical examples.     

Networked predictive control of constrained linear systems with input quantisation

This article investigates the problem of stabilising predictive control for constrained systems, wherein communication from the controller to the plant input is through a digital channel subject to quantisation and delay. A novel model with structured norm-bounded uncertainties is proposed to describe control system with input quantisation. Under a multirate scheme, a delay compensation strategy is presented. The networked predictive control synthesis approach is developed by solving a finite receding horizon optimisation problem with free control moves. It is shown that the proposed predictive controller not only efficiently reduces the negative effects of the quantisation and communication delays but also guarantees the closed-loop stability and satisfies constraints. Finally, a simulation example illustrates the effectiveness of the derived method.