基于FCM的复杂网络重叠社团结构发现算法

复杂网络中的社团结构发现是对网络数据集进行数据挖掘的普遍性问题.针对网络中大量存在的重叠社团现象,提出了基于FCM的发现重叠社团结构算法,并进一步在NG模块度的基础上,给出了评价重叠社团结构的模块度函数.算法首先将网络的节点映射成欧氏空间的节点,再以此做模糊聚类得到各重叠社团结构,根据模块度函数选择最佳重叠社团结构.最后,在经典网络上的实验结果表明,算法能够得到满意度高的重叠社团结构,而且时间复杂度较低.     

模块度优化启发式算法应用

模块度优化的启发式快速算法常常用来检测复杂网络中的社团结构.较之其余的社团检测方法,该算法在计算时间上更具优势,而且用模块度衡量发现检测社团的质量很高.运用模块度优化启发式算法划分空手道俱乐部网络、大学足球俱乐部网络和区域贸易网络等,并对其结构和功能做出一定的分析.特别地,针对贸易网络中自由贸易区往往表现为一个社团的特点,以221个国家或地区为研究对象,对贸易协定与地域之间的关系做了大量的实证研究.首先,从世贸组织网站上采集了区域贸易协定中国家之间贸易的数据;其次,通过模块度启发式算法进行社团划分,共得出7个主要的贸易区,其中欧盟自由贸易区的社团表现极为明显;最后,从社团结构的表现形式推断实际区域间的贸易情况.     

完全子图的邻域重叠社团结构探测

为了能够更准确地对邻域重叠网络进行社团结构探测,研究人员对基于完全子图的社团探测算法进行了改进。在合并完全子图团簇时,计算每一对完全子图的重叠节点个数,设置合并完全子图的阈值,如果大于阈值,则合并。在处理不在团簇内的其他节点时,采用按照比例系数大小来划分规则进行划分。算法应用5-空手道俱乐部和科学家合作网当中,验证算法可以更准确地探测邻域重叠社团结构。     

复杂网络社团结构划分方法及其应用

许多实际的复杂网络都具有明显的社团结构,整个网络的功能实现是网络中各社团相互联系作用的结果。文章分别使用Girvan-Newman算法和Newman快速算法对实际的复杂网络进行社团划分。仿真结果显示,使用两种算法对网络进行社团划分时,都能得到清晰的社团结构,使用GN算法得到的社团数目小于使用NF算法得到的社团数目。NF算法的运行速度明显比GN算法的模块度要快。     

一种基于Jaccard相似度的社团发现方法

在复杂网络中自动发现社团具有重要的实际应用价值。本文提出一种基于Jaccard相似度的社团发现方法,首先使用结点的相似度寻找优先邻居结点,再从优先邻居结点中获得最佳邻居结点,最后将互为友好结点的两个结点合并,最终得到若干个结点的组成小社团,然后利用CNM算法凝聚小社团,完成社团发现聚类。实验表明本算法执行时间缩短,所发现社团模块度增加,是一个有效的社团发现算法。     

识别稳定的局部社区结构算法

通过提取出具有高链接密度的局部团,从局部团出发进行扩展社区,设计自动选择社区扩展的终止条件,以保留最优社区结构,提出一种社会网络局部社区识别算法。在人工生成网络和真实网络上的实验结果表明,与同类算法相比,该算法能够识别出稳定的局部社区结构,提升了局部社区识别结果的准确率。     

基于用户节点相似度的局部社团挖掘

随着Web2.0的迅速发展,社交网络规模变得越来越庞大,包含的信息量也越来越多。传统的从全局角度出发进行社团挖掘的算法由于较高的复杂度以及获取全网信息的困难度,不适用于大型复杂网络的社团挖掘。文中提出了一种基于节点相似度和局部中心点的社团挖掘算法,实验结果表明,该算法在保证较高准确度的同时具有线性时间复杂度。     

基于最小熵聚类的社团检测算法

提出一种基于最小熵的社团结构检测算法。首先用模糊关系表示交互网络,一种基于熵的测度来确定模糊关系的隶属度,且熵越小,节点越相似。然后提出一种新的模糊关系合成规则,通过应用该规则,模糊关系被转换为最小关系。最后,通过熵的值把这个最小模糊关系划分成一个个社团。在人工网络与真实网络中的测试结果表明,该算法可以有效地识别社团结构。     

基于图熵聚类的重叠社区发现算法

图聚类算法是数据挖掘和复杂网络研究中的一个关键环节。基于密度、层次划分的方法已经被广泛应用于流行病学、新陈代谢和科学引文写作中。尽管上述的聚类方法适用于复杂网络的社区发现,但精度受到限制,其中一个最大的挑战是重叠社区的生成。为填补这一缺口,提出了一种利用图熵搜索局部最优的聚类方法。与传统的基于密度的种子生长式方法不同,在每一次迭代中,引入图熵来衡量图结构的模块度,并为种子的选择提供了随机选择、基于节点的度和基于节点的聚类系数3种方案。经过自下而上迭代的聚类,引入准确率和召回率等评价指标评估聚类结果的精确度,证明了算法的有效性。     

一种结合模拟退火和贪心策略的社团识别算法

为了提高复杂网络社团识别的精度和速度,文中结合模拟退火和贪心策略识别社团结构的优势,提出一种新的社团识别算法。该算法利用贪心策略引导模拟退火搜索最优解过程中单个结点的无规则盲目移动,消除了大量无效移动,在搜索到全局最优解的情况下,将搜索时间大幅缩减。实验表明,SAGA具有强大的搜索能力和较快的模拟退火执行速度,可获得较高的模块度,达到较为准确的社团分割,且具有一定的应用价值。     

Set pair three-way overlapping community discovery algorithm for weighted social internet of things

There are many problems in Social Internet of Things(IoTs), such as complex topology information, different degree of association between nodes and overlapping communities. The idea of set pair information grain computing and clustering is introduced to solve the above problems so as to accurately describe the similarity between nodes and fully explore the multi-community structure. A Set Pair Three-Way Overlapping Community Discovery Algorithm for Weighted Social Internet of Things (WSIoT-SPTOCD) is proposed. In the local network structure, which fully considers the topological information between nodes, the set pair connection degree is used to analyze the identity, difference and reverse of neighbor nodes. The similarity degree of different neighbor nodes is defined from network edge weight and node degree, and the similarity measurement method of set pair between nodes based on the local information structure is proposed. According to the number of nodes' neighbors and the connection degree of adjacent edges, the clustering intensity of nodes is defined, and an improved algorithm for initial value selection of k-means is proposed. The nodes are allocated according to the set pair similarity between nodes and different communities. Three-way community structures composed of a positive domain, boundary domain and negative domain are generated iteratively. Next, the overlapping node set is generated according to the calculation results of community node membership. Finally, experiments are carried out on artificial networks and real networks. The results show that WSIoT-SPTOCD performs well in terms of standardized mutual information, overlapping community modularity and F1.     

基于局部探测的快速复杂网络聚类算法

目前复杂网络的规模越来越庞大,且呈现天然的分布式特性,因此从局部观点出发提出快速网络聚类算法就成为迫切需要.为解决这一问题,本文基于对网络模块性函数Q的分析,推导出一个针对于单个结点的局部目标函数f,并证明Q函数随网络中任一结点的f函数呈单调递增趋势,进而提出一个基于局部优化的近线性网络聚类算法FNCA.在该算法中,每...     

一种基于边界节点识别的复杂网络局部社区发现算法

在网络日益巨大化和复杂化的背景下,挖掘全局网络的社区结构代价较高。因此,基于给定节点的局部社区发现对研究复杂网络社区结构有重要的应用意义。现有算法往往存在着稳定性和准确性不高,预设定阈值难以获取等问题。该文提出一种基于边界节点识别的复杂网络局部社区发现算法,全面比较待合并节点的连接相似性进行节点聚类;并通过边界节点识别控制局部社区的规模和范围,从而获取给定节点所属社区的完整信息。在计算机生成网络和真实网络上的实验和分析证明,该算法能够自主挖掘给定节点所属的局部社区结构,有效地提升局部社区发现稳定性和准确率。     

基于传染病模型的LPA特征阀值社团划分方法

社团结构划分对于分析复杂网络的统计特性非常重要.在非均匀社交网络的信息传播中,社团结构划分更是一个广泛关注的研究热点,相关研究往往侧重于研究紧密连接的社团结构对于信息传播所产生的关键影响.传统社团划分方法大多基于点和边的相关特性进行构建,如标签传播算法LPA（Label Propagation Algorithm）通过半监督机器学习方法,基于网络节点标签的智能交换和社团融合过程进行社团划分,但运行效率较低.为提高LPA类算法的运行速度,使其快速收敛,并提高社团划分精度,特别是重叠社团划分精度,针对LPA算法划分中的低运行效率和低融合收敛速度,本文从标签传播的网络连接矩阵本质出发,将该矩阵的最大非零特征值与网络标签信息传播的阀值相结合,提出了新的基于传染病传播模型的社团划分方法（简称ESLPA算法,Epidemic Spreading LPA）.通过经典LFR Benchmark模拟测试网络、随机网络以及真实社交网络数据上的算法验证,结果表明该算法时间复杂度大幅优于经典LPA算法,在重叠社团划分上精确度优于基于LPA模型的经典COPRA算法,特别是在重叠社团较明显时,划分精度接近精度较高GA、N-cut和A-cut算法,明显优于GN、FastGN和CPM等经典算法.     

Analysing social behaviour and message dissemination in human based delay tolerant network

Recent advances in mobile communication shows proliferation in networks formed by human carried devices known as the pocket switched network (PSN). Human beings are social animals. They tend to form groups and communities, and have repetitive mobility pattern which can be used to disseminate information in PSNs. In this paper, we give a deeper insight to the nature of community formation and how such information can be used to help opportunistic forwarding in mobile opportunistic networks. Using real world mobility traces, we first derive the adjacency list for each node and form the contact graph. Using tools from social network analysis we then determine various node properties like centrality and clustering coefficient and graph properties like average path length and modularity. Based on the derived graph properties, node encounter process and nature of message dissemination in PSNs, we propose two social based routing, known as the contact based routing and community aware two-hop routing. We compare the proposed routing techniques with generic epidemic and prophet routing and Bubble-Rap, a social based routing. Results show that the proposed algorithms is able to achieve better delivery ratio and lower delay than Bubble Rap, while reducing the high overhead ratio of epidemic and prophet routing.     

Optimization and Application of Clustering Algorithm in Community Discovery

In order to make up for the defect that the traditional spectral clustering algorithm cannot determine the number of clusters and the time-consuming calculation, this paper studies and improves the spectral clustering algorithm. In complex community networks, the spectral clustering algorithm based on modularity optimization is chosen to find the number of communities. In addition, four types of user attribute information are integrated, and a more reasonable user similarity model is constructed. At the same time, the original non-parallelized spectral clustering algorithm is optimized, and its improved scheme is suitable for the application of distributed computing. Many Hadoop optimization strategies are proposed for virtual community discovery scenarios in large-scale communities. Finally, the experimental results show that the efficiency of the parallelized spectral clustering algorithm is greatly improved, which can be applied to the virtual community discovery in large-scale social networks.     

Finding Best Matching Community for Common Nodes in Mobile Social Networks

The increase of mobile data users has created traffic congestion in current cellular networks. Due to this, mobile network providers have been facing difficulty in delivering the best services for customers. Since, detecting community in mobile social network is a valuable technique to leverage the downlink traffic congestion by enhancing local communications within the community, it attracts the attention of many researchers. Therefore, developing an algorithm, which detects community, plays a key role in mobile social network. In this paper, first, we proposed external density metrics to detect mobile social network. External density is defined as the ratio of outgoing links to total links of the community. Second, method to find the best group for common node is proposed. Therefore, an external density algorithm, makes a fair partition by grouping common nodes to a community with relatively higher external density. As a result, the overall modularity value of the network has increased. Third, the proposed algorithm is evaluated. Hence, the evaluation results confirm that our proposed approach has demonstrated good performance improvements than traditional methods.

     

Research on label propagation algorithm based on modularity maximization in the social network

A kind of community detection method based on the combination of modularity and community structure attributes was proposed.Firstly,updating the whole network after communities merging every time could result in the high time complexity,therefore,introducing propagation distance parameter and “merger going after label propagation” was utilized to reduce time complexity.Secondly,CDMM-LPA algorithm was proposed by combing label propagation with community structure.Finally,empirical analysis on data networks verified the validity of the approaches.The experimental results show that the CDMM-LPA algorithm has a high modularity value and a more stable community structure while reducing the time complexity.