A Study of Adaptive-layout for Social Network Graph
|School||Harbin Institute of Technology|
|Course||Computer Science and Technology|
|Keywords||query expansionsocial network multi level layoutm community partion adaptive layout|
Along with the development social network, kinds of social networkingwebsites appear on the Internet. People become to using these websites to recordtheir life and work more and more frequently, especially the Micro Blog. How toextract the potential information from the large scale of network data and display theinformation we are interested in briefly, thus providing more customized service hasbecome an important research area, namely information visualization.Information visualization constructs brief social network graph and display thegraph using some layout algorithms, aiming at to display the features of topology,society and semantics. So, how to choose an appropriate layout algorithm is animportant issue. Traditional layout algorithms usually proceed from the topologyfeature and ignore the community information of the network.We first introduce some common visualization technology, and then take twolayout algorithms, single Force-directed layout algorithm and multilevel layoutalgorithm, of node-link representation as example, to analyze and find the shortcomings of these traditional layout algorithms. Then we propose a new small-worldbased multilevel layout algorithm to deal with semantic network graph with obvioussocial feature. The main contributions of this paper are as follows:(1) We improved the implementation of a former multilevel layout algorithmfor we changed the computation of edge strength of social network graph, thus toreduce the effect of an unknown parameter. At the same time, the improvement ofthe algorithm obtained a better result of graph part. This also plays a role of the baseof the following algorithm.(2) According to the characteristic of social network graph, we propose a newrefiner&layout method. We consider from community and edge relation, andpropose two object function to satisfy the target of new algorithm. We solve theobject function using the gradient descent method combining the former method ofgraph partitioning, and thus form the algorithm of our multilevel algorithm.(3)We proposed a new adaptive tree layout algorithm based on the analysis ofMicro Blog Information Visualization. Our algorithm satisfied the requirement ofutilizing the layout area reasonably, focusing on important people and switch onimportant areas.(4) We also designed and developed a Micro Blog Data Analysis System. Thesystem can be used to pre-process the blog data and provide some service to browserusers. The system integrate the functions of information retrieval, layout selection,relation display, hot words filter, brief layout, and event traverse analysis.