Font Size: a A A

Research On The Construction And Application Of Enterprise Knowledge Graph Under Open Data

Posted on:2022-04-11Degree:DoctorType:Dissertation
Country:ChinaCandidate:H Q LiFull Text:PDF
GTID:1529307040469924Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
In recent years,China has issued a series of policy documents to strengthen the public sharing of enterprise information.The government attaches more and more importance to the public sharing of enterprise information,and takes it as the mechanism guarantee of sharing and co governance by taking it as government departments,industry organizations and the public to participate in supervision.The sharing of social governance cannot be separated from the efficient use of data,especially in the open data era,a large number of data are open.Since the United States launched open data movement in 2009,open government data has spread rapidly around the world.In 2018,there were 139 countries with open government data portals.As of October 2020,China has 142 government data open portals.Enterprise open data is the focus of open government data.The transparency of enterprise information is as a mechanism guarantee for government departments,industry organizations and the public to participate in supervision and realize social co governance and sharing.Enterprise open data is an indispensable part of the realization of government governance,regulatory agencies supervision and financial institutions services.Enterprise registration and administrative licensing punishment are one of the 14 common data sets on the open platform.Commercial entities are the top ten high-quality data sets with high number of records and fields.However,different from the vigorous development of open data portal,China has not launched large-scale application of open data which is in the unilateral hot situation of open data without consumption.Information users often only grasp one aspect of the information of the enterprise,unable to effectively aggregate all kinds of enterprise information and have no insight into the whole picture of the enterprise.The concept of data fusion provided by knowledge graph provides technical support for the realization of enterprise information aggregation,and can depict the complete picture of the subject concerned so as to promote data application into a new era.Therefore,this thesis has carried out a series of research on the EKG,including:(1)By investigating the problems existing in the open data application research,linked data research and the current enterprise knowledge graph construction,combined with the data sources,this thesis puts forward the enterprise knowledge graph life cycle model based on the "open data + Web resources",guiding the construction,open sharing and application of enterprise knowledge graph.(2)According to the characteristics of enterprise open data of China’s local government data open platform: most of the platforms provide open data in tabular format,the same type of data sets on different platforms adopt different schema definitions,the information coverage of the platform is not enough to support the construction of enterprise knowledge graph,and web resources are also needed,this thesis proposes the building method of enterprise knowledge graph driven by open data.Ontology is used as the form of knowledge representation.Semantic annotation is used to realize the semantic transformation from tabular to knowledge graph.Data collation and knowledge fusion are used to complete the integration of open data and web resources,and knowledge storage is used to realize the construction of enterprise knowledge graph based on graph database.In view of the current situation that ontology is difficult to be found and reused,ontology construction is guided by FAIR principle,focusing on solving the data governance problem of open data and enterprise data application problem,reusing technical standards(metadata standard,quality standard,provenance standard and Schema.org)and domain standards(Beneficial Ownership Standard,Open Contract Data Standard and e Xtensible Business Reporting Language).Based on the evaluation results,knowledge reasoning,knowledge completion and knowledge updating are used to improve the quality of enterprise knowledge graph,improve the integrity and timeliness of enterprise knowledge graph,and realize dynamic enterprise knowledge graph.(3)Taking knowledge graph as an independent new management object,guided by FAIR principle,the semantic description model,release model and provenance model of open enterprise knowledge graph are established.According to the semantic description model and release model,ontology release,data release and knowledge graph release are realized.Provenance management realizes the whole life cycle management of enterprise knowledge graph from data source generation to knowledge graph generation and then to knowledge graph reuse.The semantic description model designs metadata scheme around the description objects(knowledge graph,data set and ontology)of knowledge graph,and establishes semantic mapping with schema vocabulary to improve the findability of enterprise knowledge graph.Data release focuses on two aspects: first,the quality of the released linked data.By establishing links with external resources such as linked open data cloud,this thesis improves the linked data to reach the five-star standard and solves the problem of low data association degree of enterprises.The second aspect is the way of publishing linked data.At present,most of the published linked data are "only readable,not writable",which limits the development potential of linked data.This thesis proposes a model of enterprise readable and writable linked data release.Based on the ontology of enterprise knowledge graph,the application model of enterprise readable and writable linked data release is constructed.The enterprise readable and writable linked data release is realized through the linked data platform,which provides data support for semantic aggregation based on linked data and collaborative work of data maintenance parties,enhances the semantic potential of the constructed enterprise knowledge graph.The release of enterprise knowledge graph realizes the open sharing of enterprise knowledge graph through Open KG.On this basis,the human and machine reference format of enterprise knowledge graph is proposed to improve the findability and reusability of enterprise knowledge graph.(4)This thesis uses "graph database + linked data" to realize the application of enterprise knowledge graph and tap the potential of enterprise knowledge graph.Cypher query based on graph database provides information users with enterprise knowledge graph analysis,including: beneficial ownership graph,investment graph,enterprise shortest path graph,family graph and identification of enterprise actual controller.Aiming at the enterprise application problems of data source analysis,this thesis releases the enterprise linked data based on beneficial ownership,open contract and XBRL international standards,so as to share,combine,explore and analyze the enterprise data of different countries,gain insight into the association network,complex ownership and control structure among enterprises,and solve the problems of international anti money laundering,anti fraud,fuzzy ownership chain,open contracts international sharing,expand the international applicability of knowledge,and better integrate into the international data ecosystem.Semantic aggregation based on linked data realizes the effective aggregation of enterprise multi-source heterogeneous information,provides information users with multi-indicator aggregation of interested enterprises and their associated companies,and provides intelligent recommendation based on semantic similarity to help information users realize efficient decision-making.Using the collaborative work advantage of the linked data platform,the data producers can maintain the enterprise data cooperatively to ensure the accuracy and timeliness of the knowledge graph.The research work of this thesis completes the construction of enterprise knowledge graph by using open data and web resources.The enterprise knowledge graph is used in the application scenarios of providing services for the government,regulatory agencies,financial institutions,enterprises and investors.It provides an effective solution for further playing the role of enterprise open data in social co governance and sharing.It also provides some reference for other fields to construct knowledge graph with open data and promotes the application of open data.
Keywords/Search Tags:Enterprise Knowledge Graph, Open Data, Metadata, Linked Data, Listed Companies
PDF Full Text Request
Related items