Transcript 投影片 - Syslab

WebQuery: Searching and Visualizing the
Web through Connectivity
Jeromy Carriere, Nortel
Rick Kazman, Software Engineering Institute
元智資工所 系統實驗室 楊錫謦
2000/1/5
Outline
• Introduction
• System Description
• Visualization of Results
• Discussion
• Future Work & Conclusion
Introduction
 Finding information located somewhere on the
WWW is frequently a daunting task.
 Method:
• Yellow-pages
• content-based search tools
 Problems of the techniques above:
• vocabulary problem
• size of result
Introduction(Cont.)
 One type of information we can use to tame the
problems is :
• People form communities on the Web and
reference each other.
System Description
 Preprocessing phase
• The connectivity information is collected
 Run-time phase
• The result are fed into the VANISH tool.
Visualization of Result
 “Rick Kazman”
Visualization of Result(Cont.)
 “software engineering AND software architecture”
Visualization of Result(Cont.)
 “library”
Visualization of Result(Cont.)
 “back pain”
Discussion
 to visualize large hit set – Cone Tree
• clutters
 to draw the user’s attension to the highly ranked
nodes – bulleyes & springs-and-weights algorithms
• The springs-and-weights algorithms are expensive.
Discussion(Cont.)
 WebQuery Weaknesses
• The result may contain nodes that are highly
interlinked but represent a single repository of
information.
• WebQuery has no knowledge of aliasing of
Web sites or nodes.
Future Work & Conclusion
 Future Work:
• Incorporation into the visualization such as the
size of nodes, the correlation between keywords
and nodes.
• a mechanism for aggregation of nodes in the
visualization.
 WebQuery is powerful for searching the Web
based on connectivity and content.
Future Work & Conclusion(Cont.)
 Replace those visualization techniques with
“Core Tree”:
• How to deal with focus changing?
• Loop Links