Research On Big Data Analysis Of Scientific And Technical Data Based On Relational Graph

With the development of computer technology,the network data in the "Internet +" era was exploded.In the process of scientific and technological innovation and technological management reform,various of scientific and technological departments have accumulated such as application,review,and the process management of scientific and technological projects.These data include both structured data based on metadata and non-structural data such as declarations.Data mining and depth utilization have been the research hotspots in recent years,especially the mining of relationships between data is more conducive to expanding the value of data.Taking the selection of review experts in the process of scientific and technological management as an example,when extracting data,it is not only necessary to consider the information of the expert's field,title,unit,etc.,but also to avoid project cooperation,results cooperation,and the same unit,former colleagues and other relationship.For the various types of relationships contained in the data of scientific and technological data,how to identify,obtain and analyze the relationship between them is the key problem to be solved.Currently,many data analysis systems are based on relational databases.Relational databases have many shortcomings in relationship analysis,including low execution efficiency,complex algorithm design,and so on.Therefore,based on the widely used relational database-based data analysis system,this paper introduces the relational graph,and uses the advantages of the graph database based on Euler graph theory to solve the problem of poor correlation analysis in relational databases.Based on the relational database and graph database,this paper focuses on solving the following problems: 1.The problem of entity identification in the process of extracting source data information;2.The problem of repeated comparison of entities and attributes during the update of graph data;3.For data analysis requirements,solving the algorithm and application problems of the two analysis requirements of expert extraction and team relationship analysis in the process of graph data analysis.Based on the actual needs and the above-mentioned problems to be solved,this paper builds on the existing relational database-based data analysis and combines the introduced graph database.The main contributions are as follows: 1.An entity extraction strategy based on word frequency is proposed to extract information;2.To propose an entity similarity comparison algorithm for conflicts in the entity update process;3.To propose two types of graph data analysis methods,including repeated attribute comparison and association relationship analysis;4.On the basis of the above contributions,combined with the Hadoop big data platform,an analysis system of scientific and technological data based on relational graph was designed and implemented.The design process and implementation method used in this topic have certain value for the study of such problems.