University of Konstanz
Graduiertenkolleg / PhD Program
Computer and Information Science

Graduation Talks

title

Visual Text Feature Analysis - Analyzing text without reading it

speaker

Daniela Oelke, University Konstanz
Konstanz, Germany

date & place

Wednesday, 28.05.2008, 17:00 h
Room C252

abstract

Books, newspaper articles, patents, service reports, ... - large amounts of information are only available in textual form. Often these valuable resources are not used, because reading and analyzing those texts would take too much effort. With the help of a combination of automatic and visual techniques analytic tasks can be supported and significantly speeded up. Our goal is to develop approaches that enable the user to analyze a set of documents with respect to a certain analysis question without reading them. To achieve that goal, combinations of low-level features will be used to approximate semantic properties of the text. In the talk visualization techniques for text analysis that is based on statistical, syntactical and structural features will be presented and the planned future work on text analysis that is based on quasi-semantic text properties will be introduced.