Performance Optimization Of Interactive Visual Analysis Of Large-scale Graph Data

It is a focused issue how to analyze and mine the valuable information effectively hiding the graph data has in the field of graph-related fields.In this paper,combining with the existing technologies in the field of data analysis,we design and implement an efficient graph analysis system which is called GASys,which is based on the incremental calculating and sampling technology.And we also design and developed a series of stratified graph sampling algorithms based on the existing sampling algorithms.It smooths the biased issues on the sampling process for the state-of-art algorithms.In the statistics,sampling is an effectively method on a large-scale data.The off-the-shelf graph sampling algorithms are declined to draw the high degree or low degree nodes in the complex networks because of scale-free.The Scale-free means that degrees of different nodes are subject to a power law distribution(Zipf).So,there is a significant difference in the degrees between all the nodes.In this paper,we propose a concept of the approximate degree distribution and devise a stratified strategy using it in the complex networks.We also design a family of stratified graph sampling methods based on the stratified strategy that generalize across the full sampling algorithms based on the node,edge,and topology-based.The experimental results show that our sampling algorithms preserve several properties of different graphs and behave more accurately than other algorithms on three real-world datasets.Further,we prove the proposed algorithms are superior to the off-the-shelf algorithms in terms of the unbiasedness of the degrees and more efficient than state-of-the-art FFS and ES-i algorithms.It is a general methods for solving the biased issues on the state-of-art algorithms.The existing graph analysis system almost do not support to analyze a large-scale graph data.We analyze the problems existing in the current graph analysis system,based on incremental pro-calculating and sampling,the GASys is designed and implemented.It allows users to analyze and visualize graph data interactively and effectively.It uses data caching technology to achieve incremental loading of data,which improves the interactive performance of machine.The GASys implements the non-full-scale analysis by using sampling technology,which proves that the analysis process are faster and more accuracy.The GASys includes a large number of graph path algorithms,graph sampling algorithms,graph clustering algorithms,and graph layout algorithms.,which results in it is fitness to the different tasks.