Parallel Coordinates

Multidimensional Visualization with Applications to Data Mining and Geometric ProblemsIntroductory Course Sept. 1 – Sept. 26, 2008Prof. Heejo Lee organizer Prof. Alfred Inselberg presenterTA : Inhwan Kim neutrino37 (at) Korea.ac. krNovel skills for discovering complex relations in multivariate data and uncovering insights into multidimensional problems are acquired. The mathematical foundations of Parallel Coordinates are intuitively presented showing how multidimensional relations are accurately mapped into 2-dimensional patterns.In turn, properties of the corresponding relation are visualized and powerful geometrical algorithms suitable for applications (e.g. data mining, process control, linear-programming, computer vision, geometric modeling and others) are generated.Software(Parallax) for data exploration and classication will be distributed to the participants. Duringthe Friday laboratory sessions we will study interactively:  visual data mining (USA Patent) on real datasets (i.e. feature extraction, GIS, financial, processcontrol, and others with many variables) - bring your multivariate datasets for analysis,  collision avoidance algorithms for air traffic control- 3 USA Patents,  detecting coplanarity, near-coplanarity, several coplanar clusters and proximities - USA Patent,  geometric classier providing the explicit classication rule and minimal set of variables (features),  nonlinear visual models from data for interactive decision-support, constraint and trade-off analyses.KEYWORDS: Multidimensional Visualization, Parallel Coordinates, Visual Data Mining and Automatic Classification, Collision Avoidance for Air Traffic Control, Non-linear Models