Research And Design Of Data Link Intelligent Analysis System Based On Knowledge Graph

In order to achieve the strategic goal of "three types and two networks,world-class",the State Grid Corporation of China is currently working to build a data center with visible data,mature components and standardized systems.The key task is to build a full-link system for data to achieve the data connection between the headquarters,provinces and cities.Due to the large scale of the data system,and the data transfer links between business systems have the characteristics of multiple links and long periods,the data links still have problems such as low data transmission effectiveness,difficulty in locating link fault,and lack of link fault warning mechanisms.In response to the above problems,the main research content is to firstly propose the PSO-XGBoost model to realize the current fault classification prediction of the data link,and then design the SWRL reasoning rule based on the Jaccard coefficient to construct the knowledge graph of the data link fault domain to realize the data link fault warning,and finally design and implement a data link intelligent analysis system based on the knowledge graph.The main research work includes the following three aspects:Firstly,propose a PSO-XGBoost data link fault classification model.The SG-UEP longitudinal link data is analyzed and the data characteristics are explored.Select the Extreme Gradient Boosting(XGBoost)algorithm as the basic model and use the particle swarm optimization(PSO)algorithm to optimizes the important parameters in the XGBoost model,and construct the PSO-XGBoost fault classification model.Through the experimental comparison with the Support Vector Machine(SVM)model,the BP neural network model and the normal XGBoost model,the effectiveness of the PSO-XGBoost model is demonstrated.The classification diagnosis results of the PSO-XGBoost model provides the current data link fault information for subsequent associated fault warnings.Secondly,design the Jaccard coefficients reasoning rule and construct the knowledge graph of data link fault domain.Using the advantages of knowledge graphs in association analysis,the ontology modeling tool Protege is used to build the ontology model of the data link fault domain and create examples.Analyze the similarity between faults based on the Jaccard coefficient and construct Semantic Web Rule Language(SWRL)rules library.Through the combination of ontology reasoning mechanism and SWRL rule description,Jena reasoning engine is used to perform knowledge reasoning on the relationship between the faults that has been established.Finally,the graphical representation ability of the graph database Neo4j is used to construct a visual knowledge graph of data link fault to achieve data link fault early warning and to improve the operation efficiency and maintenance efficiency of the data link.Thirdly,design and implement a data link intelligent analysis system based on knowledge graph.Based on the above research results,analyze the system requirements,design the system architecture,form the machine learning fault classification module and the knowledge graph fault warning module,and realize the data link intelligent analysis system based on the knowledge graph.