用户知识模型的自动建立和演进技术_论文.docVIP

摘 要 为了解决搜索引擎查准率的问题，个性化检索技术已逐渐成为搜索引擎研究 的热点之一，以期能够更有效、更准确地找到用户感兴趣的内容。 本文针对个性化信息检索系统的核心问题用户模型开展研究。首先介绍了 知识管理、搜索引擎和用户模型及相关反馈的发展历史、现状和发展趋势，然后 对信息检索和用户模型及相关反馈的理论和技术进行综述，最后指出目前存在的 主要问题。 本文通过对用户兴趣模型和领域知识模型的分析，提出用户知识模型。该模 型综合考虑了用户知识、用户兴趣所包含的信息，将用户模型作为知识组织与存 储与个性化信息检索作为知识利用分离开来，并把相关反馈作为知识获取的一种 手段，其所包含的信息不仅比传统信息检索技术所包含的资源信息更为丰富，而 且包含的用户信息更为丰富，并且能识别用户新的知识和兴趣，能为用户发现新 的感兴趣的文档。文章讨论其定义、性质、计算方法等建模理论，使用粗糙概念 格表示用户知识模型，给出建模和演进算法，完成用户知识模型自动建立和演进。 本文最后讨论在 J2EE 平台上设计开发了个性化信息检索系统石鼓搜索， 分析功能模块、系统设计和原型开发过程，原型系统已成功应用于南华大学门户 网站，取得较好效果。 关键词：个性化搜索引擎 用户知识模型 粗糙概念格 相关反馈 I Abstract For precision of search engine, personalized information retrieval has been the focus question of research to search out one’s most interested contents precisely and efficiently. This paper focuses on core of personalized search engine - user model. Firstly, the paper sum up developing history, status, tendency of knowledge management, search engine, user model and relevant feedback, then introduces the main issues of information retrieval, user model, relevant feedback theories and technology, points out the main questions of existing state at last. After analyzing user interest model and domain knowledge model, user knowledge model is brought forward in this paper, which thinks over the information contained in the user knowledge, the user interest model. It is sepreated user knowledge model as knowledge organization from personalized information retrieval as knowledge application, and the user relevant feedback is regarded as one of methods in knowledge acquisition. The model contains more information than the resource information of traditional information retrieval model and the user interest information of traditional information model, and moreover, the rough concept lattice based personalized information retrieval model can identify the new knowledge and interest of users and find the documents of new interest. In this paper we discussed theory include concept, properties and algorithm of user knowledge model, pr