Research On Semantic Quality Evaluation Model Of Event Knowledge Graph Based On DQV

In recent years,with the integration of technologies such as big data,artificial intelligence and the Internet into various fields,the digital economy has become a key force for global development and competition.As a special data resource,the quality of Event Knowledge Graph(EKG)has become a difficulty and pain point in current development.Especially in the oil industry,it also focused on EKG’s data quality problems.At present,most scholars mainly focus on the construction and application level of event knowledge graphs.However,the quality problems such as semantic inconsistency,semantic redundancy,and semantic incompleteness in the constructed event knowledge graphs affect the application and sharing of event knowledge graph data.As an important tool for event knowledge graph quality management and evaluation,the quality model aims to provide standard terminology and measurement function guidance.However,the current existing quality models are inconsistent in terminology,difficult to expand and inapplicable,resulting in certain difficulties in the quality management and evaluation process of event knowledge graphs,making it difficult to share and reuse data in event knowledge graphs.In view of the above problems,this thesis firstly studies the Data Quality Vocabulary(DQV)and other four classical quality models,and unifies the problem of term inconsistency in quality models.Secondly,the current existing quality model only provides users with a reference model and does not systematically propose a set of quality model modeling methods,which leads to vague concept and use of quality models in the process of EKG quality management and evaluation by users.Therefore,it is necessary to systematically propose a set of quality model modeling methods to provide guidance for users.The key problems studied and solved are as follows:1.Aiming at the problem of inconsistent terminology in the quality model,this thesis sorted out and summarized the DQV standard vocabulary and other four classical quality models through the traceability method,extracted the common core elements of the quality model and constructed it as the data quality conceptual model,which laid the foundation for the subsequent construction of the terminology unification of the semantic quality model of the event knowledge graph.2.Aiming at the difficulty of quality model construction,this thesis proposes a set of Three-stage Six-step Methodology for Quality Modeling(TS＿MQM)based on the GQM idea under the premise of unifying quality model terminology.TS＿MQM includes three stages: concept modeling,vocabulary mapping,and model building.The concept modeling stage is the process of refining the concept of the quality model on the basis of unifying terms.In the vocabulary mapping stage,the principle of interoperability should be followed,and the reuse of the DQV standard vocabulary should be preferred.The model construction phase is to combine the actual business requirements of EKG to finally build a standard,consistent and applicable event knowledge graph semantic quality model.3.In order to improve the EKG quality management and evaluation tasks,four important EKG semantic quality dimensions are established from the actual business needs of EKG under the guidance of TS＿MQM method.Based on the DQV standard vocabulary and combined with the TS＿MQM method,the Event Semantic Quality Vocabulary(ESQV)is constructed from the bottom up.The constructed ESQV provides standard terminology and measurement function guidance for EKG quality management and evaluation.Therefore,the EKG semantic quality measurement function is designed and implemented based on the hybrid method under the guidance of ESQV.4.Finally,taking the event knowledge graph in the field of downhole operations as an example,the semantic quality evaluation system of the event knowledge graph is designed and implemented using Python,Django and other technologies,which verifies the validity and feasibility of the research content of this thesis.Through the above research,this thesis aims to provide a standard,normative and referable quality model and evaluation framework for event knowledge graph quality management and evaluation,and provides a method for the quality evaluation of event knowledge graph data.