摘要
针对企业现有招投标文档价值信息挖掘不足、文档知识难以应用等问题,设计一种基于知识图谱的招标项目文档智能管理系统。系统核心功能模块包括项目管理、模板管理、知识图谱和统计查询。项目管理和模板管理模块分别对项目文档进行分类管理和提供知识抽取模板。知识图谱模块实现文档知识抽取,并将抽取的知识与元数据构建知识图谱,实现文档的语义互联。对于文档知识抽取分别提出预训练模型结合规则配置的文字知识抽取模型和图片分类与光学字符识别融合的图片知识提取模型。统计查询模块基于构建的文档知识图谱实现多维统计分析、语义检索与智能问答等应用。该文档管理系统以智能化技术支持文档知识深度挖掘和反馈,能够实现文档价值充分利用。
Aiming at the problems of insufficient value information mining and difficult knowledge application of existing en⁃terprise project documents,an intelligent project project document management system based on knowledge graph is designed.The system mainly includes four functional modules:project management,template management,knowledge graph and statistical query.The project and template management modules are to classify and manage project documents and provide knowledge extraction tem⁃plates.The knowledge graph module realizes document knowledge extraction and constructs a knowledge graph with the extracted knowledge and metadata to achieve the semantic interconnection of documents.For document knowledge extraction,a text knowl⁃edge extraction model based on a pre⁃training model and rule design and a picture knowledge extraction model combined with pic⁃ture classification and optical character recognition are proposed.The statistical query module supports multidimensional statistical analysis,semantic retrieval and intelligent question answering applications based on the constructed knowledge graph.The docu⁃ment management system supports the deep mining and feedback of document knowledge with intelligent technologies and can make full use of document value.
作者
王志刚
吴士泓
李孟全
李向
Wang Zhigang;Wu Shihong;Li Mengquan;Li Xiang(YGSoft Inc.,Zhuhai 519085;School of Management,Huazhong University of Science and Technology,Wuhan 430074)
出处
《现代计算机》
2023年第3期111-120,共10页
Modern Computer
关键词
文档智能管理系统
知识图谱
自然语言处理
光学字符识别
机器学习
智能问答
document intelligent management system
knowledge graph
neural language processing
optical character recog⁃nition
machine learning
intelligent question answering