Visualization of Relational Structure Among Scientific Articles

This paper describes an ongoing technique to collecting, mining, clustering and visualizing scientific articles and their relations in information science. We aim to provide a valuable tool for researchers in quick analyzing the relationship and retrieving the relevant documents. Our system, called CAVis, first automatically searches and retrieves articles from the Internet using given keywords. These articles are next converted into readable text documents. The system next analyzes these documents and it creates similarity matrix. A clustering algorithm is then applied to group the relevant papers into corresponding clusters. Finally, we provide a visual interface so that users can easily view the structure and the citing relations among articles. From the view, they can navigate through the collection as well as retrieve a particular article.