【作者】 刘星焜；

【导师】 陈鄞；

【作者基本信息】 哈尔滨工业大学 ， 软件工程， 2020， 硕士

【摘要】 随着服务计算、云计算、物联网、大数据等高新技术的快速发展,由现实世界与虚拟网络空间之间各类线上线下的服务互联互通,产生了大服务的现象。大服务源于互联网中多样化、异构和复杂的海量服务,是由异构服务聚合而来的,因此大服务平台需要整合大量的分散异构服务资源,这些服务资源的获取,存储和管理存在以下难点:服务资源的来源众多、服务资源的设计没有统一的模式,因此数据充满多样性和异构性、服务资源有多种使用方式,需要有通用的操作接口对服务资源进行调用,因此,在此背景下本文使用知识图谱帮助解决以上问题,提出基于众包方式构建知识图谱的方法,众包引入的模板进行对齐和融合得到统一模板的方法,使用模板对服务资源进行统一描述的方法,并基于以上方法设计和实现了面向大服务的知识图谱管理平台。平台为用户提供具体领域的知识图谱构建服务,降低用户构建知识图谱的门槛;支持用户根据自己的服务资源格式,自定义知识图谱的模板;不同用户的知识图谱进行独立存储;采用众包方式获取服务资源,增量的构建知识图谱;提供通用的知识图谱查询与基于图嵌入模型的推理应用接口,支持第三方应用程序基于上述接口进行二次开发。平台的知识图谱构建关键技术包括基于众包的知识图谱构建方法,基于OWL的知识图谱模板定义与对齐和融合方法以及异构服务资源的语义对齐和融合方法。在全面进行需求分析和知识图谱构建关键技术分析的基础上,本文将平台划分为图空间管理模块、模板模块、资源模块、图嵌入模块和后台管理模块五个部分,并对各个模块进行设计和实现。最终本文按照需求分析,对系统进行总体设计和详细设计,对面向大服务的知识图谱管理平台进行了系统实现并对系统进行测试。更多还原

【Abstract】 With the rapid development of high technology such as service computing,cloud computing,Io T,and big data,various types of online and offline services are interconnected between real-world and virtual network space,resulting in the phenomenon of big service.Big service originates from the diversified,heterogeneous,and complex services on the Internet,which are aggregated from heterogeneous services.Therefore,the big service platform needs to integrate a large number of dispersed heterogeneous service resources.The following difficulties exist in the acquisition,storage,and management of these service resources: there are many sources of service resources,there is no unified model for the design of service resources,so the data is full of diversity and heterogeneity,and there are many ways to use service resources,and general operation interfaces are needed to use the resources.Therefore,this paper uses knowledge graphs to help solve the above problems.We propose a method of constructing knowledge graphs based on crowdsourcing.The schema introduced by crowdsourcing is aligned and merged to obtain a unified schema.We use the schema to describe the service resources in a unified way.We designed and implemented a big-service-oriented knowledge graph management platform based on the above methods.The platform provides users with knowledge graph construction services,simplifies the construction of knowledge graph;supports users to customize knowledge graph schema according to their resource;stores different users’ knowledge graphs independently;uses the idea of crowdsourcing to get service resource and builds knowledge graph incrementally;provides general knowledge graph operation interfaces such as graph query interface and graph-embedding-model-based reasoning interface which can support secondary development for third-party applications.The key technologies of knowledge graph construction used by the platform include the knowledge graph construction method based on crowdsourcing,the definition of OWL-based knowledge graph schema and its aligning and merging method,the aligning and merging method of heterogeneous service resources.Based on the comprehensive analysis of requirements and key technical analysis of knowledge graph construction,this paper divides the platform into five parts: graph space management module,schema module,resource module,graph embedding module,and background management module.Then we designed and implemented the above modules.In the end,according to the requirement analysis,the overall design and detailed design of the system are carried out.After that,we implemented and tested the platform.更多还原