【作者】 钱炜源；

【导师】 梁昌勇；

【作者基本信息】 合肥工业大学 ， 管理科学与工程， 2007， 硕士

【摘要】 在知识经济到来的今天，知识管理系统成为了企业进行知识管理的有效工具，它能有效地对企业知识进行组织、管理和应用。知识管理系统中的知识检索系统，是员工获取知识，学习知识的重要工具。因此，提高知识检索效率可以有效地提高知识学习以及运用能力。传统的知识检索系统中，不能根据不同的用户需求来提供准确的知识，查准率比较低。本文根据知识管理系统中知识库存储知识的情况以及知识检索特点，结合目前的知识管理系统和搜索引擎技术设计了个性化知识检索系统，给出了个性化知识检索系统的模型结构，对个性化知识检索中的核心技术，即用户建模技术，知识过滤技术等方面做了研究。在用户兴趣模型和知识文档模型的建模方式上，提出了一种基于特征词权重向量和知识分类概率分布相结合的模型，使得用户兴趣以及知识文档的表示更加全面；对于检索系统中的知识过滤，运用结合了基于内容过滤与协作过滤的综合过滤方式，提出了一种返回结果的个性化排序算法。通过个性化知识检索系统中各个模块的相互协作来达到在知识检索中实现个性化功能。更多还原

【Abstract】 In knowledge economy arrival today, the knowledge management system became an effective tool for the enterprise to implement the knowledge management, it could effectively carry on the organization, management and application of the enterprise knowledge. The knowledge retrieval system in the knowledge management system is an important tool for the staff to acquire and study knowledge. Therefore, enhancing the knowledge retrieval efficiency avails effective improvement of the knowledge study as well as the utilization ability.In the traditional knowledge retrieval system, it could not provide accurate knowledge according to the different user demand, so the accuracy ratio is low. In this thesis, we design a personalized knowledge retrieval system which integrates present knowledge management system and search engine technology according to the knowledge storage situation in knowledge base as well as the knowledge retrieval characteristic, propose an personalized knowledge retrieval system model structure and do research on user profile construction and knowledge filtering technology which are the core technologies in personalized knowledge retrieval. This thesis provides a model of user profile and knowledge document combining term vector space mode and knowledge classification probability model which makes the user profile and knowledge document expression more comprehensive, uses a compositive filtering way which integrates content-base filtering and collaborative filtering on knowledge filtering and propose a personalized ranking algorithm of the retrieval results. Modules in the personalized knowledge retrieval system cooperate to achieve personalization function.更多还原