The Research Of Visualization-oriented Evaluation Methods For Graph Layouts

With the arrival of information age,complex networks include social network,financial transaction network,biology network and so on,producing a large number of large-scale data every day.Data analysis and data mining for complex networks can explore a lot of meaningful information.In the complex network,the number of nodes is very large,the relationship between edges is also very complex,so traditional ways such as texts and forms can’t meet the user’s requirements of understanding the structure of network.The visual layout of network presents the underlying structure of the network intuitively,so that people can understand the relational data more deeply.Therefore,the automatic visual layout method of complex network has been the major method of analyzing the network data.There exists a large amount of network layout methods,and graph layout algorithms have different advantages and disadvantages from different aspects.The quality of graph layout greatly affects the user’s understanding of graph structure.Some visualization means are mainly proposed from a technical perspective,while ignoring the psychological cognition of human.As a consequence,many visualization results can’t be understood and accepted by users.Based on human perception,the scientific evaluation of the intuition and effectiveness of different graph layouts can help the users select the layout that conforms to their psychological recognition,so that users can see through the internal relationship of the network data.So evaluating the graph layouts based on the characteristic of a graph and users’ psychological cognitive has significantly theoretical and application value.This paper proposes two graph layout evaluation methods.Firstly,this paper proposes an evaluation method of graph layouts based on visual perception,and it is applicable to the situation when users focus on important nodes.It analyzes the nodes’visual importance based on a user experiment and designs a model to quantify the nodes’ visual importance.Then it evaluates the pros and cons of graph layouts by comparing the topological importance and visual importance of nodes.A heatmap-based visualization is used to provide visual presentation for the difference between the topological importance and visual importance of nodes.Meanwhile,a metric is built to quantify the difference precisely.Finally,experiments are done under different scale of data sets to further analyze the characteristics of these graph layout methods.The layout with small differences between visual importance and topology importance is preferred.At the same time,this paper proposes an objective overall quality assessment method for graph layout algorithms.Firstly,we build the subjective rating database of graph layouts.The subject experiment is designed to rate different graph layouts.Then,for each graph layout,we use the readability metrics of the layout as independent variables,the subjective score of users as the dependent variable,to establish the regression model.Through the regression model,we can get the overall quality score of a graph layout.Two graph layout evaluation methods in this paper have different applicable scenarios.Selecting the graph layout that conforms to the user’s psychological recognition helps users deeply understand the feature of the graph.The experiment shows that our method is in good agreement with the actual situation.