Design And Application Of Knowledge Graph Visualization Platform

With the rapid development of Internet technology,the massive data generated by various services is growing exponentially,and research on efficient analysis and mining about big data for specific scenarios has been extensive in last decade.Faced with massive amounts of data,practitioners in different industries need to face different scenarios and need to conduct quickly and analysis on massive amounts of data.As an emerging technology that combines the advantages of big data and deep learning,knowledge graphs can meet the needs of data mining and analysis for specific fields well.The knowledge graph technology provides a better ability to organize,manage and understand the massive information of the Internet,and express the big data of the Internet into another totally new form closer to the human cognitive world.However,the current existing knowledge graphs and visualization services have the following problems:1)The maintainability of the knowledge graphs is poor.It requires professionals to follow up in real time and add or delete data,which consumes lots of human resource and is correct in terms of maintainability and availability.Non-professionals are not friendly;2)Few researchers combine knowledge graphs with visualization techniques,resulting in insufficient display and insufficient utilization of knowledge graphs;3)The coupling of the knowledge graph visualization platform and data is too strong,which makes the platform difficult to reuse and migrate to other data,which greatly increases the development cost,and lacks an integrated system including data processing,knowledge graph,and visualization platform after proper decoupling;4)The knowledge graph visualization platform for mass data is in the knowledge graph There are bottlenecks and limitations in data storage and query.Most of the existing knowledge graph visualization platform solutions cannot effectively alleviate or break through this bottleneck for the time being.In response to the above problems,through systematic research on different parts of the knowledge graph visualization platform,we build a knowledge graph visualization platform to provide users in different industries with fast data analysis and mining for industry big data.This research mainly includes the following content:(1)Design an integrated platform including data processing,knowledge graph storage and visualization.This platform not only has integrated and convenient operation procedures,but can greatly simplify practitioners in specific industries.The use and maintenance process of the platform,and has a certain degree of decoupling,can be migrated to and used in other industry data under certain configuration changes;(2)Apply the platform to the massive tax data provided by partners,In response to the specific needs of big tax data,the different modules of the platform have been improved to realize the visualization of knowledge graphs based on tax big data for taxpayer relationship queries,enterprise purchase-sales relationship queries and other practical issues,especially in mass data The need for storage and complex iterative queries has achieved a breakthrough in the bottleneck;(3)The platform is applied to the massive mobile phone business record data provided by another partner,and the analysis and prediction of the user’s job and residence problems based on mobile big data is realized Predict the relationship with users,and design a set of neural network models embedded in the data processing module of the platform,which greatly improves the accuracy of specific data prediction and judgment.The knowledge graph visualization platform designed and implemented in this thesis supports the dynamic addition,deletion and modification of data,and is connected to the relevant data interface of the partner,realizing the timing and non-perceptual import of incremental data,and realizing related knowledge graphs for different business needs The design and visualization display meet the needs of different industries for rapid analysis and query of industry big data,and provide good user interaction in the visualization module.The platform well combines and utilizes emerging technologies such as distributed storage,Spark big data computing,deep learning,and knowledge graphs,and provides improved ideas and specific practices for knowledge graph visualization in various directions.