Description of DocumentAtlas

Purpose and Functionality

Software for visualization and interactive browsing of textual documents – it supports 2D view and 3D (VRML) view to the contents of visualized documents. It is based on using dimensionality reduction for document visualization by first extracting main concepts from documents and than using this information to position documents on a two dimensional plane via multidimensional scaling. The final output is graphical presentation of a document set that can be plotted on a computer screen.

Availability, Preconditions and Licensing

Installation is publicly available in binaries at http://docatlas.ijs.si/The software runs under Windows platform .NET framework 2.0.

The tool is free for research purposes.

Integration with other SEKT Tools

Developed on the top of Text-Garden tool and integrated with Text Garden, development so far inside the SEKT project.
Document Atlas is standalone tool but also integrated into OntoGen tool for semi-automatic ontology construction.

Publications

FORTUNA, Blaž, MLADENIĆ, Dunja, GROBELNIK, Marko. Visualization of text document corpus. Informatica (Ljublj.), 2005, vol. 29, no. 4, 497-502.

FORTUNA, Blaž, GROBELNIK, Marko, MLADENIĆ, Dunja. Visualization of document corpus. V: ANŽIČ, Tina (ur.), GROBELNIK, Marko (ur.), HORVAT, Boris (ur.), MLADENIĆ, Dunja (ur.), PISANSKI, Tomaž (ur.), SHAWE-TAYLOR, John (ur.), ŠKVARČ, Smiljana (ur.), ŽEROVNIK, Janez (ur.). Complex objects visualization – COV 2005 : proceedings. Ljubljana: Jožef Stefan Institute: IMFM – Institute of Mathematics, Physics and Mechanics; Koper: UP – PINT, 2006.

GROBELNIK, Marko, MLADENIĆ, Dunja. Visualizing Very Large Graphs Using Clustering Neighborhoods. V: Local Pattern Detection. Volume 3539/2005. Springer Berlin / Heidelberg.