【作者】 周小玲；

【导师】 胡于进；

【作者基本信息】 华中科技大学 ， 管理科学与工程， 2004， 硕士

【摘要】 为满足用户在网络时代的办公需求,适应知识经济时代企业管理的要求,本文基于Web应用,以知识管理为指导思想,设计开发了一个办公自动化系统。该系统基于B/S结构,能够通过Internet访问,同时系统安全也有一定的保证。工作流是办公自动化系统中的关键技术。本文在工作流组成及其模型的基础上,设计了一个流程定制模块,实现了工作流的流程定义、环节设置和流向定义。给出了一个项目申请工作流实例,实现了工作流中办公文档的传送。利用表单存取控制和文档存取控制技术加强了办公文档的安全性。信息集成体现知识共享的内容,通过OA系统能够实现企业信息集成。本文介绍了实现信息集成的关键技术,给出了以OA系统为基础的信息集成基本框架。利用Domino集成工具和LS数据对象,实现了OA系统对外部关系数据库的访问。全文检索和文本分类是文本挖掘的重要内容。本文首先对数据库创建了全文索引,同时,设计了一个Web查询表单。该表单根据用户的输入,调用确定查询的URL值,实现了基于Web的文档检索查询。本文对现有贝叶斯文本分类方法进行了分析改进,首先提取待分类文本的特征向量,建立一个向量空间模型,然后采用贝叶斯文本分类方法,利用概率统计对文档进行分类。最后,本文给出了该文本分类方法的一个测试实例。更多还原

【Abstract】 In order to meet user’s requirement, in this thesis an office automation system is researched on and developed, which is an OA system based on Web and knowledge management oriented. Based on B/S architecture, the OA system can be visited via Internet, and at the same time the system’s security is assured in some degree.The workflow is the main technology of the OA system. Firstly, the composing frame and model of workflow are introduced, and then the workflow-setting module is designed, by which the flow definition, link setting and flow-direction setting are implemented. To discussing workflow more exactly, a workflow example of item application is presented, in which the transmission of official documents is accomplished. Through controlling the allowance of forms and documents, the security of official documents is improved.The information integration of enterprise can be realized through the OA system. In this thesis, the information integration and its support technologies are introduced firstly. Then, the basic frame of information integration is provided based on OA system. With data integration tools of Domino and LS Data Object, the access of the OA system to external relation database is realized.Full-text retrieval and text classification are the important technologies in text mining. In this thesis, the full-text index of database is established and a retrieval form is designed. According to the input of user in the form, the URL value is confirmed by a key function, and the Web-based retrieval is realized. This thesis analyzes and improves the Bayesian text classification method. In details, the characteristic vectors of the documents are abstracted firstly, and then a vector space model is established, with probability statistics, the documents are classified by the Bayesian text classification method. At last, a test instance of text classification is presented.更多还原