| Summary: | 碩士 === 國立臺北大學 === 資訊管理研究所 === 95 === Information visualization can help researchers know the revolution of knowledge domain and future development in large amount of citation data. We use open source Prefuse and programming language JAVA to implement a visualization system. Our system is able to draw graphs from GraphML file, .net file and database. In order to reduce the number of unnecessary nodes, our system provides two kinds of pruning methods: in-degree filter and out-degree filter. Combined with degree filter and title keyword search, we attain keyword relative domain’s citation connection.
Compared with Pajek in several cases, our system shows better visual presentation. We can get various graphs by connecting to database, not just the single one graph represented in the graph file. Also, Our system can deal with weighted graphs, find nodes’ traversal, and search for specified nodes. From the meanings of graphs, we are able to understand the cluster relationships. Furthermore, we can research deeper theories based on these earlier works.
|
|---|
