Locally linear discriminant embedding: An efficient method for face recognition

In this paper an efficient feature extraction method named as locally linear discriminant embedding (LLDE) is proposed for face recognition. It is well known that a point can be linearly reconstructed by its neighbors and the reconstruction weights are under the sum-to-one constraint in the classical locally linear embedding (LLE). So the constrained weights obey an important symmetry: for any particular data point, they are invariant to rotations, rescalings and translations. The latter two are introduced to the proposed method to strengthen the classification ability of the original LLE. The data with different class labels are translated by the corresponding vectors and those belonging to the same class are translated by the same vector. In order to cluster the data with the same label closer, they are also rescaled to some extent. So after translation and rescaling, the discriminability of the data will be improved significantly. The proposed method is compared with some related feature extraction methods such as maximum margin criterion (MMC), as well as other supervised manifold learning-based approaches, for example ensemble unified LLE and linear discriminant analysis (En-ULLELDA), locally linear discriminant analysis (LLDA). Experimental results on Yale and CMU PIE face databases convince us that the proposed method provides a better representation of the class information and obtains much higher recognition accuracies.     

Small sample improvements in the threshold cointegration test using residual-based moving block bootstrap

A residual-based moving block bootstrap procedure for testing the null hypothesis of linear cointegration versus cointegration with threshold effects is proposed. When the regressors and errors of the models are serially and contemporaneously correlated, our test compares favourably with the Sup LM test proposed by Gonzalo and Pitarakis. Indeed, shortcomings of the former motivated the development of our test. The small sample performance of the bootstrap test is investigated by Monte Carlo simulations, and the results show that the test performs better than the Sup LM test.     

Cracking the walnut: Using a computer game to impact cognition,emotion, and behavior of highly aggressive fifth grade students

Three fifth grade highly aggressive students participated in a study examining effects of an aggression intervention game. Using single-subject multiple baseline research design, the effect of the game was measured across three domains: attribution of peer intent (cognitive), emotional response (affective) and behavior response (behavioral). The results indicate that, while there were observable individual differences, the game had an immediate impact on attribution of peer behavior across all participants, had a less predictable impact on reported behavioral reactions, and had no discernable impact on reported emotional response. Theoretical and practical implications of the study are discussed.     

利用网络编码优化Ad Hoc网安全路由协议SAODV

网络编码可大大提高网络吞吐量、减少延迟。然而,由于编码意义上网络信息流的复杂性,实际应用中仍存在困难。另一方面,Ad Hoc网作为自组织形式的特殊网络,其安全路由协议是研究重点。针对Ad Hoc网安全路由协议中较优秀的SAODV,提出一种基于网络编码的优化方案,较传统编码易于实现、具有较强实用性。     

复方降脂中药的性味归经与药效的BPNN研究

针对复方中药药效与其性味归经之间复杂的、不确定的关系,建立适于复方降脂中药的BP神经网络模型,运用Matlab快速预测复方降脂中药的药效指标。经验证,该方法为一种快速、可靠的方法。     

BPEL流程监视的可视化实现

When a BPEL process is executed, it is necessary to dynamically monitor the process. BPEL is a executable language, which is not suitable for visual monitoring. On the other hand, BPMN is designed to visually describe business process and is more intuitive for monitoring. To visually monitor a BPEL process, transformation from BPEL to BPMN is necessary. However, current study of transformation from BPEL to BPMN does not support the transformation of "link" activity. Besides, no work has been done to add supplementary information into BPMN during transformation. In this paper, we transform nested BPEL process into a flat BPMN process graph without hierarchy through applying a flattening strategy. Especially, we analyze various scenarios of the transformation of link activity, and provide a method to deal with it. Besides, we analyze the mapping between BPEL activities and BPMN graph, through which we found out that some supplementary information cannot automatically obtained from BPEL process. These supplementary information need to be added during transformation. At the end of this paper, we present the structure of our monitoring tool which is based on our transformation algorithm.     

CAD模型表面区域分割方法

三维模型表面区域分割技术在形状分析,尤其是局部形状分析中具有重要作用,传统的表面分割方法主要针对网格模型,而机械工程中的CAD模型通常用B-rep表达.提出一种B-rep形式的CAD模型表面区域分割方法,将模型表面划分为局部凸区域、凹区域和平区域的组合,并使得分割后的区域数量最少.为提高计算效率,提出一种二步法:首先在模型面的局部凸凹性分析的基础上,快速地将模型表面分割成初始的凸区域、凹区域和平区域;然后通过区域合并的方法对分割后的区域进行组合优化,得到一个最优的分割结果.实验结果证明,该方法能有效地分割模型的表面区域.     

基于UML扩展的数据仓库多维数据PIM建模

针对数据仓库多维数据PIM建模问题,对多维数据模型维和事实进行了定义.在UML(Unified Modeling Language)扩展的基础上,提出了与多维模型相关的八个版类,从而可以对数据仓库中PIM模型的各种情况进行图形化描述.最后基于某钢铁企业采购系统的采购价格差异分析主题进行了建模.     

利用POMDP模型来增强分布式系统的生存性

针对分布式系统存在的状态信息不完全问题,引入部分可观察的马尔可夫决策过程(POMDP)模型到生存控制系统中.在该控制系统的构造过程中,结合前瞻的思想,提出一种简易、有效的搜索算法(NSL算法)来作出决策,从而在一定程度上弥补了现有生存控制系统的不足,提高了分布式系统的可生存性.     

基于移动Agent的电子商务交易平台研究

针对传统电子商务平台存在的问题,提出了一个以移动Agent为中介服务的电子商务交易平台,阐述了该平台各组成部分的功能.实验采用价格优先和数量优先的购买策略,分析了该平台中移动Agent的性能,实验仿真结果表明移动Agent的引入提高了平台处理客户信息的速度,缩短了达成交易的时间并降低了网络的通信量.