We have just released our program VisuMap 1.4. The standard edition of the
program is free for non-commercial use. It can be downloaded at:
VisuMap is a visualization tool for high dimensional data. It basically
turns any numerical table into a 2D map that preserves, as much as possible,
the distance information of the original dataset. VisuMap is based on a new
dimensionality reducing method called relational perspective map (RPM). RPM
works like traditional multidimensional scaling (MDS) methods, but shows
quite different characteristics.
Additionally VisuMap also implements other MDS methods like Sammon mapping,
curvilinear component analysise, principal component analysis as well as
clustering methods like K-mean clustering and self-organizing map (i.e.
The following link is an example that depicts DOW 30 index stocks. Each spot
on the map represents a stock. Closely located stocks have similar weekly
price history in the year 2002. The price histograms of some stocks are
depicted as examples.
New features and major changes in VisuMap 1.4
Implemented a new optimization algorithm for RPM map that is about
10 times faster than previous simulated annealing based algorithm
Implemented two new mapping algorithms: Sammon mapping and the
curvilinear component analysis.
Many enhancements in user interfaces: configurable toolbars,
drag-and-drop supports, customizable glyphs etc.
Upgraded the platform from .NET 1.0 to .NET 1.1.
James X. Li