RWO-Sampling: A random walk over-sampling approach to imbalanced data classification

This study investigates how to alleviate the class imbalance problems for constructing unbiased classifiers when instances in one class are more than that in another. Since keeping the data distribution unchanged and expanding class boundaries after synthetic samples have been added influence the classification performance greatly, we take into account the above two factors, and propose a Random Walk Over-Sampling approach (RWO-Sampling) to balancing different class samples by creating synthetic samples through randomly walking from the real data. When some conditions are satisfied, it can be proved that, both the expected average and the standard deviation of the generated samples equal to that of the original minority class data. RWO-Sampling also expands the minority class boundary after synthetic samples have been generated. In this work, we perform a broad experimental evaluation, and experimental results show that, RWO-Sampling statistically does much better than alternative methods on imbalanced data sets when implementing common baseline algorithms.     

LBSN协作式个性化链接预测算法

在基于位置的社交网络中用户链接与位置链接之间具有一定的内在关联,而且不同的用户在社交网络中的表现也存在差异,因此对于以上问题提出一种协作式个性化链接预测算法。针对用户的个性化特征,采用核密度估计方式对用户在时间和空间维度建模,基于兴趣组对用户进行重叠社团划分,并通过社团、好友以及签到关系进行个性化用户链接预测;基于个性化用户链接预测结果,利用从社团重启的随机游走预测用户的个性化位置链接;协作式个性化链接预测算法通过用户链接预测和位置链接预测的迭代使得两者性能相互提升。实验结果表明,所提算法相比于现有算法具有更好的预测性能。     

基于随机游走的改进标签传播算法

社区发现是挖掘社交网络隐藏信息的一个有用的工具,而标签传播算法（LPA）是社区发现算法中的一种常见算法,不需要任何的先验知识,且运行速度快。针对标签传播算法有很强的随机性而导致的社区发现算法结果不稳定的问题,提出了一种基于随机游走的改进标签传播算法（LPARW）。首先,根据在网络上进行随机游走确定了节点重要性的排序,从而得到节点的更新顺序;然后,遍历节点的更新序列,对每个节点将其与排序在其之前的节点进行相似性计算,若该节点与排序在其之前的节点是邻居节点且它们之间的相似性大于阈值,则将排序在其之前的节点选为种子节点;最后,将种子节点的标签传播给其余的节点,得到社区的最终划分结果。将所提算法与一些经典的标签传播算法在4个有标签的网络和5个无标签的真实网络上进行比较分析,实验结果表明所提算法在标准互信息（NMI）、调整兰德系数（ARI）和模块度等经典的评价指标上的性能均优于其余对比算法,可见该算法具有很好的社区划分效果。     

A Highly Secured Image Encryption Scheme using Quantum Walk and Chaos

The use of multimedia data sharing has drastically increased in the past few decades due to the revolutionary improvements in communication technologies such as the 4th generation (4G) and 5th generation (5G) etc. Researchers have proposed many image encryption algorithms based on the classical random walk and chaos theory for sharing an image in a secure way. Instead of the classical random walk, this paper proposes the quantum walk to achieve high image security. Classical random walk exhibits randomness due to the stochastic transitions between states, on the other hand, the quantum walk is more random and achieve randomness due to the superposition, and the interference of the wave functions. The proposed image encryption scheme is evaluated using extensive security metrics such as correlation coefficient, entropy, histogram, time complexity, number of pixels change rate and unified average intensity etc. All experimental results validate the proposed scheme, and it is concluded that the proposed scheme is highly secured, lightweight and computationally efficient. In the proposed scheme, the values of the correlation coefficient, entropy, mean square error (MSE), number of pixels change rate (NPCR), unified average change intensity (UACI) and contrast are 0.0069, 7.9970, 40.39, 99.60%, 33.47 and 10.4542 respectively.     

一种高效的压缩Page Walk Cache结构

通用图形处理单元（GPGPU）已被广泛应用于现代高性能计算系统中。GPGPU的单指令多线程执行模型导致快表命中率较低,特别是对于那些不规则应用,需要借助PWC减少实际的页表访问次数。传统PWC中存在很多冗余信息,加之容量有限,实际效果并不理想。分析了传统PWC中的信息冗余情况,提出了一种新结构——压缩PWC。压缩PWC在保证查找开销不变的基础上,完全消除了冗余信息,压缩了空间,使PWC能够记录更多的页表访问历史,从而有效减少地址转换过程中访问页表的次数。测试结果表明,与相同容量的传统PWC相比,压缩PWC可以显著缩短虚实地址转换时间开销。     

基于量子漫步算法的地震震前异常挖掘

地震特别是大震前会产生一些异常,但这些异常信息难以识别,导致无法充分利用这些异常信息预测地震的发生时间,减少地震带来的灾害影响.针对这个问题,提出一种基于量子漫步算法的震前异常挖掘方法,提取汶川地震和芦山地震的震前射出长波辐射（Outgoing Long-wave Radiation,OLR）异常,进而计算地震前后的P值,异常值CD等数据,通过统计分析方法,探索OLR异常与地震的关系.并且通过实验将该算法扩展到最近十年左右全球发生的8.0级及以上地震,验证该算法的有效性.实验结果表明,该算法能够有效的反映在地震前后会出现OLR异常,而且越大的地震异常越明显.因此,该算法适用于震前异常挖掘.     

基于马尔科夫随机游走的两阶段离群检测算法

基于邻域的离群点检测算法中,参数的选择与确定是一个重要的问题,不合理的参数选择导致算法的性能显著下降。为减少参数对于离群点检测的影响,提出了一种基于马尔科夫随机游走的两阶段离群检测算法,可以在不影响算法效率的基础上,有效降低参数对检测结果的影响。该算法采用均匀采样策略生成一系列三角剖分图,并引入移除规则得到节点的拓扑结构,从而获得由节点连通性定义的转移概率矩阵,有效减少了算法的计算量和运行时间;其采用加权投票原则重新定义重启向量,并将不同图上得到的平稳分布向量的平均偏差值作为离群点分数,有效地提高了算法的准确性。采用合成数据集以及UCI数据集,验证了该算法与现有的算法相比有更高的准确率。     

Modified progressive random walk with chaotic PRNG

Abstract

In this paper, two modifications are proposed to the Progressive Random Walk (PRW) algorithm in order to address its potentially insufficient search space coverage. The first modification replaces the Pseudo-Random Number Generator (PRNG) with the uniform distribution by the chaotic map based PRNG for generating of the offset values and the second modification is called direction switching and is based on experiment observation. The modifications are implemented into the PRW and the resulting algorithm is called modified Progressive Random Walk. The search space coverage of the two algorithms is compared. Both algorithms are used in macro ruggedness estimation of the CEC2015 benchmark set and the results are discussed.     

Relationship Identification Across Heterogeneous Online Social Networks

In the era of the social web, many people manage their social relationships through various online social networking services. It has been found that identifying the types of social relationships among users in online social networks facilitates the marketing of products via electronic “word of mouth.” However, it is a great challenge to identify the types of social relationships, given very limited information in a social network. In this article, we study how to identify the types of relationships across multiple heterogeneous social networks and examine if combining certain information from different social networks can help improve the identification accuracy. The main contribution of our research is that we develop a novel decision tree initiated random walk model, which takes into account both global network structure and local user behavior to bootstrap the performance of relationship identification. Experiments conducted based on two real‐world social networks, Sina Weibo and Jiepang, demonstrate that the proposed model achieves an average accuracy of 92.0%, significantly outperforming other baseline methods. Our experiments also confirm the effectiveness of combining information from multiple social networks. Moreover, our results reveal that human mobility features indicating location categories, coincidence, and check‐in patterns are among the most discriminative features for relationship identification.     

基于图论的运动对象分割

将两种基于图论的算法图切割与随机游走应用于运动对象的分割。利用图切割,通过构造网络图、HSV空间消除阴影、建立混合模型背景及运动区域最小切割自动完成运动对象的分割;利用随机游走,通过选择种子点、建立权函数、求解Dirichlet问题及形态滤波等交互地完成运动对象的分割。通过实验对两种算法的应用进行了研究。结果表明:基于图切割算法可针对单目标快速、有效地进行分割,并获得干净、光滑的分割结果;基于随机游走算法可针对微弱物体边界或低对比度的图像分割,可在不要求实时性的情况下获取更为精确的分割结果,为后续的跟踪、自动分割等处理提供基础。