Font Size: a A A

Enterprise Data Environment Construction Combined With Semantic Linked Open Data

Posted on:2018-12-07Degree:DoctorType:Dissertation
Country:ChinaCandidate:C XieFull Text:PDF
GTID:1368330590455303Subject:Software engineering
Abstract/Summary:PDF Full Text Request
With the rapid growth of the complexity of business in large scale and group enterprises,the construction of large enterprise data environment is facing the following two challenges:(1)difficult in integrating and sharing complex business data.Business entities integration is rather difficult due to the huge amount of entities,complex relationships among the entities and frequently change in large enterprise information system.For example,in a typical group of manufacturing enterprises,the size of the business entity from thousands to tens of thousands of different,but the relationship between entities is close to the ten times the number of entities.The modification or adjustment of the business entity will lead to the difficulty of the entity information sharing in the complex business.(2)Difficult in deeply analysing on business data.Due to the lack of semantic information in traditional enterprise data environment,it is difficult to analyze the enterprise data deeply.a large number of clinical data For example,due to the lack of drug side effects,alternative medicine and other fields of knowledge support;lack of disease knowledge in medical records;lack of support of in additional information of treatments and surgical operations;It is hardly to deeply analyze the hidden relationships between the in medical data environment.The lack of semantic information is one of the main reasons that hinder the data integration,knowledge sharing and deep analysis in current enterprises.The current studies suggest,based on existing data environment and open knowledge,to build a semantic data environment for complex data sharing and analysis is a feasible path to resolve the problems.In this paper,we propose a semantic data environment construction framework based on the Linked Model,Linked Open Data,and semantic matching.The paper tries to combine the open knowledge data with the internal data of the enterprise,to improve the ability of knowledge sharing and data semantic analysis in the traditional enterprises.It aims to provide a possible way to realize the intellectualization in today's enterprise data environment.The framework is applied to a largescale production enterprises and a large central hospital.The results verify the feasibility and effectiveness of the framework.In details,the main contributions the paper are the following: In this paper,we first propose a complete enterprise data semantic support framework,which is combined with an open environment to build a rich semantic enterprise data application environment.The framework uses the traditional database and other data sources as input,and outputs an enterprise semantic data environment.The framework includes three parts: data modeling,knowledge import and semantic data fusion,which can support complex business sharing and deeply business data analysing. In this paper,we propose an analysis-orient data model that mapping the traditional relational schema into the linked data model.It is able to map the internal data of a relational database to the linked data described in RDF.Meanwhile,it could transform the semi-structured form described by the E-R relation model into the Linked Data described by RDF.The general Linked Data transformation is applied on a manufacturing enterprise as well as database transformation is applied on a central hospital.It provides the base data model for both the two enterprise and other applications in data analysis. In this paper,we propose a complete Linked Data linking method to import the external knowledge into the internal data environment.It not only can be used to link the instances in the open data,but also can be used to link high-level conceptual objects through instance mining and description reasoning.The linking of instances is applied to the database of drugs in a hospital,which greatly enriches the dimensions of drug information in HIS system.The concept linking is applied to a wheel manufacturing enterprise,which greatly enriches the knowledge interpretation of the professional process concept,and shortens the gap between designing and field processing. In this paper,we propose a semantic matching method based on the linked data graph,which can integrate the knowledge data and the internal data automatically.It combines the semantic matching techniques,such as Sim Rank,Sim Flooding,into a complete matching algorithm.The algorithm can effectively deal with the semantic matching in large-scale relational data graphs such as productions,materials,parts in a manufacturing production or drugs,treatments,disease in a hospital.The algorithm has a good result in the benchmark test of OAEI.The algorithm has also been applied to discover similar cases in a hospital and provided a new and effective solution for medical data analysis. In this paper,a complete data semantic support framework has been applied to large scale production and service enterprises.Based on specific business requirements,importing and combining open knowledge,the framework effectively improves the enterprise's production process document creation,dissemination,adjustment and on-site deployment,shorten the gaps between designing and field processing of the manufacturing enterprise.By analyzing semantic labels from data semantic support framework,a central hospital could accurately identify scalpers out of normal users in mobile medical service.In general,according to different business requirements,the proposed the framework can be applied flexibly in different business scenarios to solve practical problems,to provide an effective knowledge sharing and data analysis support environment for modern production and service enterprises.It also provides a feasible way for the modern enterprise to implement intellectualization in data management.
Keywords/Search Tags:Semantic Web, Linked Data, Linked Open Data, Semantic Matching, Semantic Fusion, Resource Service
PDF Full Text Request
Related items