 |
 |
|||||||||||||
 |
 |
 |
 |
 |
 |
 |
|
 |
|
 |
|
 |
|
|
 |
||||||||||||||
|
|
An important technological aspect of knowledge management is to support companies in their efforts to analyze their text corpora, i.e. to structure, to condensate and to learn from documented knowledge assets. Information retrieval systems offer a content-based and query-driven (and therefore goal-directed) search in document collections. However, goal-directed search assumes that the user has a pretty good understanding of what he is looking for. In contrast, for analyzing documented knowledge, explorative search plays an important role. Here, the focus of interest is rather blurred and may cover wide areas of a collection’s topics. The need for semantic search paths arises which provide insight into the semantic structure of a corpus of documents. This
work studies the concept of a graphical presentation of a text
collection’s similarity structure: It presents an application-oriented
design and evaluation of a document map approach for visually aiding
analysis tasks which are relevant for managing the knowledge contained
in specialized document collections. An open framework allows an
application-specific adaptation of the method for generating such a
graphical corpus overview. Based upon this framework the interactive
document map system DocMINER comprises tools for analyzing specialized
text collections. Industrial and scientific case studies help to better
understand what analysis tasks document maps support, and whether they
can be effectively applied to real-world problems. Moreover, a
comparative empirical study in a laboratory setting evaluates the
document map concept against query-driven retrieval interfaces. The
contributions of this work can be considered from two complementary
points of view: From the perspective of the application field of
knowledge management the thesis provides advanced tool support for
the important and difficult problem of exploiting documented knowledge.
Furthermore, the work shows meaningful and useful application scenarios
for document maps, proposes methods for their fruitful use in these
contexts, and presents experience reports from which other real-world
projects concerned with structuring, condensing and exploiting
documented knowledge can benefit. From the perspective of information
visualization the work studies a visual metaphor for a clearly
defined application domain and contributes to the understanding of the
analysis tasks this metaphor supports. This work shows the strengths and
weaknesses, the application potential and the limits of the
visualization method and thus may help to improve existing and to
develop new approaches for graphically displaying complex textual
information. Links
PublicationsContactDr. Andreas Becks Fraunhofer Institute for Applied Information Technology Office:
Informatik V, RWTH Aachen, Ahornstr. 55 |
|||||||||||||||||
