语义对等网环境下基于用户兴趣的检索系统模型分析-model analysis of retrieval system based on user interest in semantic peer-to-peer network environment.docxVIP

摘要 摘要 语义对等网是一种结合了语义网技术和 P2P 技术的新型对等网络。该网络中 每个节点的资源可以通过统一明确的模式描述，从而利用机器自动处理这些资源， 充分发挥对等网模式下信息表示的潜力。如何有效利用该环境下的语义信息，提 高信息资源的检索效率已经得到越来越多学者的关注，成为了新的研究热点。针 对现有研究中仅从浅层次方面利用语义对等网中的语义信息，忽视语义推理机制 和语义关系，本文提出利用本地节点用户兴趣来构建检索系统模型。 本文首先对国内外学者的研究现状进行概述，总结当前研究方法的不足。然 后，以语义对等网为基础研究平台，利用知识地图在信息组织与管理上的优势构 建本地用户兴趣模型。同时应用语义匹配技术，使具有相似兴趣的节点根据其之 间的内容关联度及兴趣相似度建立邻接关系，并依照相应规则形成用户兴趣组。 最后，结合已形成的用户兴趣组提出基于用户兴趣的信息检索模型，深入论述该 系统各模块的功能，指出所采用的关键技术，并说明该系统实现信息资源查找的 动态过程。 通过对语义对等网环境下基于用户兴趣检索系统模型的研究，极大提高了资 源的利用率与检索的准确性，使节点用户获得满意的信息资源，从而摆脱原有语 义对等网检索效率低下，资源利用不全的弊端。 关键词：语义对等网 用户兴趣模型 知识地图 检索模型 Ab Abstract Abstract Semantic peer-to-peer web which is combined with the semantic web technology and P2P technology provides a new pattern web. In this web, the machine can automatically process the resources through the unified pattern and extract implicit semantic information. Apparently, the semantic peer-to-peer web is capable to enrich the semantic representation of the peer-processed information. How to effectively use the semantic information in this environment and improve the efficiency of searching information resources has become catch more and more attention of scholars. Considering the method for information retrieval in e xisting research mainly used semantic equivalence of shallow levels and neglected semantic reasoning mechanism and semantic relations, a construction method of retrieval model is proposed in this paper. Firstly, it scientifically organized various methods for information management and retrieval in semantic peer-to-peer web, summarized the shortcomings of current research. Then, with the usage of the advantages of knowledge map on the information organization and management, the local peer interest model can be constructed. Meanwhile, it applied the semantic matching technology to establish the adjacency relations on the basis of content as well as the interest similarity degree involving two peers and form user interest group in accordance with the rules of corresponding. Finally an