28.06.2013 Views

Papers in PDF format

Papers in PDF format

Papers in PDF format

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

3D Visualization of WWW Semantic Content for Brows<strong>in</strong>g and Query<br />

Formulation<br />

1. Introduction<br />

Richard H. Fowler, Wendy A. L. Fowler, and Jorge L. Williams<br />

Department of Computer Science<br />

University of Texas - Pan American<br />

Ed<strong>in</strong>burg, TX, USA 78539-2999<br />

email: {fowler, wfowler, jwilliams}@panam.edu<br />

Abstract: Visualization is a promis<strong>in</strong>g technique for both enhanc<strong>in</strong>g users’ perception of<br />

structure <strong>in</strong> the Internet and provid<strong>in</strong>g navigation facilities for its large <strong>in</strong><strong>format</strong>ion spaces.<br />

This paper describes an application of the Document Explorer to the visualization of WWW<br />

content structure. The system provides visualization, brows<strong>in</strong>g, and query formulation<br />

mechanisms based on documents’ semantic content. These mechanisms complement text and<br />

l<strong>in</strong>k based search by supply<strong>in</strong>g a visual search and query formulation environment us<strong>in</strong>g semantic<br />

associations among documents. The user can view and <strong>in</strong>teract with visual representations<br />

of WWW document relations to traverse this derived document space. The relationships<br />

among <strong>in</strong>dividual keywords <strong>in</strong> the documents are also represented visually to support<br />

query formulation by direct manipulation of content words <strong>in</strong> the document set. A suite of<br />

tools is provided for navigation and orientation <strong>in</strong> the visual network representations of<br />

document set and term collection.<br />

The explosive growth of the Internet has added to the need to filter and organize <strong>in</strong><strong>format</strong>ion so that users<br />

can efficiently and effectively identify relevant documents. Browsers such as Mosaic [Schatz & Hard<strong>in</strong>, 1994]<br />

for the World Wide Web and Harmony [Andrews & Kappe, 1994] for Hyper-G allow users to traverse the <strong>in</strong><strong>format</strong>ion<br />

space of documents through connections provided by document authors. Though this may account<br />

for much of the popularity of the Internet, it hides many of the difficulties of effective <strong>in</strong><strong>format</strong>ion access.<br />

Search<strong>in</strong>g Internet resources is difficult due to size, diversity of data, and lack of a common <strong>in</strong>dex<strong>in</strong>g scheme.<br />

The most widely used search tools are automatically generated search systems like Lycos and Harvest and<br />

manually organized systems like Yahoo and Internet Yellow pages. Though such tools are valuable assets to<br />

the Internet searcher, it seems likely that these tools alone will not solve the current problems of <strong>in</strong><strong>format</strong>ion<br />

access.<br />

The challenges of <strong>in</strong><strong>format</strong>ion access on the Internet are issues common to all forms of <strong>in</strong><strong>format</strong>ion retrieval.<br />

These issues <strong>in</strong>clude difficulties <strong>in</strong> us<strong>in</strong>g <strong>in</strong>dex<strong>in</strong>g vocabularies, <strong>in</strong>dex<strong>in</strong>g <strong>in</strong>determ<strong>in</strong>acy, and the<br />

user’s <strong>in</strong>ability to completely specify <strong>in</strong><strong>format</strong>ion needs [Ingwerson & Wormell, 1986]. Retriev<strong>in</strong>g <strong>in</strong><strong>format</strong>ion<br />

that meets users’ <strong>in</strong><strong>format</strong>ion needs is an iterative process, and techniques which explicitly <strong>in</strong>corporate users’<br />

judgments, such as relevance feedback [Maron & Kuhn, 1960], provide means to automate some aspects of<br />

user guided retrieval. It is also clear that mechanisms provid<strong>in</strong>g alternative paths of access to <strong>in</strong><strong>format</strong>ion can<br />

enhance retrieval effectiveness [Bates, 1986].<br />

One promis<strong>in</strong>g approach for enhanc<strong>in</strong>g <strong>in</strong><strong>format</strong>ion retrieval through the Internet is visualization to facilitate<br />

users’ perception of document relation structure. A number of systems have been developed to provide<br />

visually based brows<strong>in</strong>g mechanisms for travers<strong>in</strong>g the l<strong>in</strong>k structure of Internet documents. McCahill and<br />

Erickson [McCahill & Erickson, 1995] describe designs for three-dimensional spatial <strong>in</strong>terfaces for Internet<br />

Gopher which employ icons with various shapes and textures to represent documents arranged by file structure<br />

or search results. The Harmony browser for Hyper-G [Andrews, 1995] provides two-dimensional structure<br />

maps of a document’s l<strong>in</strong>k neighborhood, as well the three-dimensional In<strong>format</strong>ion Landscape visualization<br />

of l<strong>in</strong>k structure hierarchy. The Narcissus system [Hendley et al., 1995] creates a three-dimensional visual<br />

representation of WWW l<strong>in</strong>k connectivity with documents’ positions determ<strong>in</strong>ed by user manipulable attractive<br />

and repulsive forces among document nodes. Munzner and Burchard [Munzner & Burchard, 1996] describe a<br />

system which displays WWW documents positioned <strong>in</strong> a hyperbolic space to <strong>in</strong>crease display density compared<br />

to conventional display spaces.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!