【作者】 何少锋；

【导师】 王邦；

【作者基本信息】 华中科技大学 ， 电子与通信工程， 2020， 硕士

【摘要】 知识图谱数据建模能够打破不同应用场景下的数据隔离,为搜索、推荐、问题、解释与决策等应用提供基础支撑。现有的知识图谱工具软件的功能存在较为片面,大多数都是提供数据分析的功能,在其软件实现中缺少独立的知识图谱建模与数据存储的功能。针对以上问题,在非全日制工程硕士两年间,本文首次提出设计并实现一个基于前后端分离开发的知识图谱建模在线应用服务系统。本系统架构采用浏览器和服务器框架,主要功能包括知识图谱建模、图谱数据存储、图谱可视化、图谱数据搜索和图谱融合。知识图谱建模功能包括图谱数据模型建模和三元组数据建模,图谱数据模型以本体集合、实体集合、关系基础和属性基础组成。图谱数据存储功能模块为了解决节点和关系属性内容不固定的问题,本文采用了静态属性与动态属性结合的技术方式将属性内容存储在Neo4j图数据库中。图谱可视化功能提供通用数据可视化能力,并且可进行自动化布局。图谱数据搜索功能可进行节点搜索、关系搜索、三元组搜索和属性搜索。图谱融合功能模块在图谱融合上首次进行了探索与实现,包括图谱信息融合、节点融合和关系融合。最后,作者对应用系统进行了功能完整性测试、运行兼容性测试和系统性能测试,测试表明本图谱建模系统功能较为齐全且运行良好,是知识图谱工具方面一次良好的探索与实现。本系统的应用功能还有较大的提升空间,在未来将提供更为优质的服务,包括速度更快容量更大的图数据存储、大规模的图谱数据可视化和智能化的图谱融合功能。更多还原

【Abstract】 Knowledge graph data modeling can break the data isolation in different application scenarios and provide basic support for applications such as search,recommendation,problem,interpretation and decision-making.The functions of existing knowledge graph tool software are relatively one-sided.Most of them provide data analysis functions,and lack of independent knowledge graph modeling and data storage functions in their software implementation.In response to the above problems,in the two-year part-time engineering master’s degree,this paper proposes for the first time to design and implement an online application service system based on knowledge graph modeling based on separate front-end and back-end development.This system architecture uses a browser and server framework.Its main functions include knowledge graph modeling,graph data storage,graph visualization,graph data search,and graph fusion.The function of knowledge graph modeling includes graph data model modeling and triple data modeling.The graph data model is composed of ontology collection,entity collection,relationship basis and attribute basis.In order to solve the problem that the content of nodes and relational attributes is not fixed,the graph data storage function module uses a combination of static attributes and dynamic attributes to store the attribute content in the Neo4 j graph database.Atlas visualization provides general data visualization capabilities and can be automated for layout.The graph data search function can perform node search,relation search,triple search and attribute search.The graph fusion function module explores and implements graph fusion for the first time,including graph information fusion,node fusion,and relationship fusion.Finally,the author performed functional integrity test,operation compatibility test,and system performance test on the application system.The tests show that this graph modeling system is relatively complete and runs well.There is still much room for improvement in the application functions of this system.In the future,it will provide better services,including faster and larger map data storage,large-scale map data visualization and intelligent map fusion functions.更多还原