导　　师： 傅秀芬

学科专业： H1203

授予学位： 硕士

作　　者： ;

机构地区： 广东工业大学

摘　　要： 随着网络资源的不断增加，人们在信息海洋中获取需要的信息也变得越来越困难，出现了“信息迷航”和“信息过载”的现象。人们急需高性能的检索技术来满足日益增长的信息需求，这促使信息检索技术的快速发展，成为当前研究热点之一。 信息检索模型是影响一个信息检索系统性能的最主要因素，通过结合用户相关反馈来提高检索系统的性能一直是众多学者研究的课题，其主要的驱动思想是高效的检索性能必定依赖于正确的用户反馈。但靠单用户进行反馈对系统检索性能的改善是有限的，利用多用户协作的方式进行反馈就显得更合理，能构造出更完善的查询请求来提高检索性能。 论文首先分析了信息检索技术的研究现状，探讨了cscw的三要素和关键技术，研究了三种典型的信息检索模型、用户相关反馈技术及搜索结果合并算法，针对现有信息检索存在的精度低、效率差和覆盖率低等问题，结合cscw、用户相关反馈和元搜索技术，提出了协同反馈信息检索模型。在此基础上，论文对基于cscw的信息检索系统进行构架，该系统融入了层次化结构的设计思想，以基于消息原语的通信机制为协同工作的基础，将协同反馈信息检索模型结合到元搜索引擎中，实现了多用户协同信息检索。最后，本文实现了系统原型的核心模块，解决了协同反馈信息检索技术中的部分问题，并通过实验验证了所提模型的可行性和有效性，以期对信息检索技术的研究提供参考和帮助。 With the increasing of Internet resouces, it becomes more and more difficult for people to obtain the information they needed, and this cause the phenomenon of 'getting lost in information' and 'information overload'. People urgently need high-performance information retrieval technology /(IRT/) to meet the growing demand for information, that results in the rapid development of the IRT and make it become one of the research hot spots. Very many factors can affect the performance of an information retrieval system /(IRS/), and the most important is the information retrieval model. How to unify with user relevant feedback technology to improve the performance of an IRS has been the research topic of many scholars, and its main thought is the highly effective information retrieval surely relies on the correct user feedback. But depending on the single user to get the correct feedback has a lot of limitation to the improvement of performance, while using the multi-user collaborative work method to get the cerrect feedback is more reasonable, and it is easier to construct good query expression to improve the performance. The paper firstly analyzed the present researching status of the IRT, discussed three essential factors and key technologies of CSCW, and researched three classic information retrieval models, user relevance feedback technology and result merging algorithms. Secondly, aiming at existing problems of IR, such as low accuracy, low rates of information coverage and bad efficiency etc, combined with CSCW, user relevance feedback technology and meta-search technology, we put forward a collaborative feedback information retrieval model. On this basis, the paper carried on the analysis and the research to the design of collaborative IRS, and discussed the framework of IRS based on the CSCW. This system take message primitive based communication mechanism as the foundation of collaborative work, applies collaborative feedback information retrieval model to the meta search engine, and has realized the multi-user collaborative information retrieval function.Finally, this paper implemented system's core modules, solved some problems related with collaborative feedback and IRT, and confirmed the feasibility and the validity of the model with a view to provide reference and assistance to the research of IRT.

关 键 词： 信息检索 检索模型 用户相关反馈 协同反馈

分 类 号： [TP391]

领　　域： [自动化与计算机技术] [自动化与计算机技术]