The Study On Complex Relational Data Visualization Technology

Visualization of information and visualization of large data focus on the visual display of large-scale non-numerical information resources to help people understand and analyze data.The effective mining of massive relational data based on the method of data mining,clustering algorithm and force-directed layout algorithm,proposes a new clustering algorithm based on data visualization method.According to the relationship between the data sets do simple statistics and comparative analysis,and then through the clustering algorithm on the cluster,will be the main node and weight,the relationship between data through the force directed algorithm,write elastic graph layout,finally based on graph visualization criteria to draw the basic framework of the layout.Through echarts and D3.js as data visualization tools and reasonable quality evaluation,the accuracy rate of 10000 data cases is 89%,which proves the rapidity and accuracy of the proposed algorithm.The thesis mainly focus on issues as follows:(1)Study and contrast between visualization algorithms used commonly in relational network visualization field,and then make a detailed classification among them based on aesthetic standard and visualization performance.(2)Evaluate the performance and results of different kinds of clustering algorithms in data mining field(3)Combine the theory of semi-supervised learning and multi-relationship data analysis,thus,giving the definition of an improved k-means clustering algorithm.(4)Improve the FR layout algorithm by compressing the relational graph with the spanning tree algorithm.(5)Through testing,the visualization processes this thesis suggested is proved to be valid to more precisely visualizes the massive relational data.