在线社会网络中融合多源信任信息的信任管理研究

项目名称： 在线社会网络中融合多源信任信息的信任管理研究

项目编号： No.61202119

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 计算机科学学科

项目作者： 常俊胜

作者单位： 中国人民解放军国防科学技术大学

项目金额： 24万元

中文摘要： 在线社会网络近年来发展迅猛，给人们生活和工作带来极大便利，但同时也给人类和信息系统的信任关系带来巨大挑战。面对大规模动态开放的在线社会网络，现有的信任管理研究工作存在对信誉数据稀疏问题重视不够、信誉系统冷启动、信任评价的信息源单一等问题。对此，项目组创新性地提出了融合多源信任信息的信任管理，通过有效融合在线社会网络中丰富的多源信任信息，包括直接经验、社区内的推荐、相关社区的信誉信息、用户提供的信任凭证及系统挖掘的相关信任信息，可以实现更为合理准确的信任评价。具体研究内容包括：保护隐私和提供激励相结合的信誉共享机制、基于本体映射的跨社区信誉转换机制、基于模糊理论的融合多源信任信息的信任模型、融合多源信任信息的信任管理原型系统。本项目的研究成果为在线社会网络中信任及信誉研究提供了全新思路与可行方案，对提高信任及信誉机制效能、增强在线社会网络可信性具有重要意义。

中文关键词： 在线社会网络；信任；信任管理；信誉；信任模型

英文摘要： The rapid development of online social networks has make people's work and life more convenient and efficient. Howerver, it causes a huge challenge for the trust between human and information systems. Trust management in large-scale, dynamic and open online social networks is currently faced with some serious problems, such as ignoring the reputation data sparsity, reputation system cold start, less trust information source for trust evaluation. Hence we make the first attempt to propose trust management combining trust information from multi-sources, which can perform trust valuation reasonably and acutely by combing plenty of trust information from multi-sources in online social networks, including direct experience, recommendations from self community, recommendations from related communities, trust credentials provided by users, associated trust information mined and etc. The research contents mainly include reputation sharing mechanism that combines privacy protection and incentive provision, cross-community reputation transform mechanism that is based on ontology mapping, fuzzy theory based trust model that combines trust information from multi-sources and trust management prototype systems. The result of our research project provides a new approach and feasible solution for current trust and reputation re

英文关键词： online social networks；trust；trust management；reputation；trust model