隐匿社区中基于缺失链接修复的重要节点发现与追踪研究

项目名称： 隐匿社区中基于缺失链接修复的重要节点发现与追踪研究

项目编号： No.61272422

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 张伟

作者单位： 南京邮电大学

项目金额： 81万元

中文摘要： 随着社会网络的发展，出现了一种以特殊形态存在的网络社区- - 隐匿社区，利用隐匿社区进行有效而隐蔽的通信成为新的安全问题。本课题以社会网络中的隐匿社区的演化为研究对象，对在大规模的社会网络中挖掘、跟踪和分析隐匿社区的组织结构和行为模式进行科学研究。具体内容包括：从社会学角度分析隐匿社区的自然属性，从工程学的角度分析社会网络中隐匿社区存在和运行的现状；研究在不确定信息条件下构建隐匿社区模型的算法，结合隐匿社区内部的信任机制，使用信任推理方法发现信任关系；设计缺失链接修复和潜在重要节点发现的算法，提出隐匿社区以其重要节点为主线的演化模型；通过开发隐匿社区模型测试床，实现隐匿社区发现和跟踪原型并加以测试。本课题旨在揭示隐匿社区及其重要节点的演化规律，将在计算科学、智能数据处理、网络安全等基础研究方面取得进展，对采用计算思维解决网络安全相关问题的解决模式产生重要影响，推进社会计算的广泛深入研究。

中文关键词： 社区发现；影响力扩散模型；网络预测；模糊系统；差分隐私

英文摘要： With the development of social networks, there has been a special form of online community - hidden community. Effective and covert communication hidden in community becomes the new security issues. This research will take the community evolution in social network as the object and study the structural organization and behavior patterns in hidden community from the view of mining, tracking and analysis. Specific contents are as follows: With the sociological view to analyze the natural attributes and the engineering view to analyze the existence and operation in hidden community; Study the algorithm which build a hidden community model with uncertain conditions, then use trust reasoning to found the trust relationships combined with the trust mechanism; Design the missing links repair and latent core nodes discovery algorithm and come up with an evolutionary model using its core nodes as the main clue; Realize the hidden community detection and tracking prototype through developing a test bed of the hidden community and then test it. Our project aims to reveal the evolution law of the hidden community and its core nodes. It will make progress in basic research fields, such as computational science, intelligent data processing, network security, and make an important impact on the solving model that uses computa

英文关键词： Community Detection；Influence Diffusion Model；Network Prediction；Fuzzy System；Differential Privacy