Conference 3 Dec - 6 Dec Exhibition 4 Dec - 6 Dec



    01 Full Conference 1 - Full Conference One Day

    Parallel Coordinates are Better than … they Look!

    Wednesday, 03 December

    14:15 - 18:00

    Rose Hall 3

    The goal of visualization is to systematically incorporate our fantastic pattern-recognition ability into the problem-solving process. With parallel coordinates the perceptual barrier imposed by our 3-dimensional habitation is breached enabling the visualization of multidimensional problems. By learning to discover patterns, corresponding to relations, from the displays a powerful knowledge discovery process (USA patent) has evolved. It is illustrated on real multivariate datasets together with guidelines for exploration and good query design; relational patterns are more compact and informative that than the data itself. Realizing that this approach is intrinsically limited leads to a deeper geometrical insight, the recognition of M-dimensional objects recursively from their (M−1)-dimensional subsets. Powerful geometrical algorithms (for intersections, containment, proximities) as well as applications including classification emerge. Surfaces in 3-D (and higher) are represented by planar regions containing information about the surfaces’ normal vectors. This yields stunning patterns unlocking new geometrical insights. Non-convexities like folds, bumps, coiling, dimples and more are no longer hidden and are detected from just one orientation. Evidently this representation is preferable for some applications like ray-tracing even in 3-D. Such results were first discovered visually in 3-D and then proved mathematically for higher dimensions; in the true spirit of Geometry! The patterns persist in the presence of errors (USA patent) and that’s good news for the applications. Applications to collision avoidance for air traffic control (3 USA patents), non-linear interactive models of complex systems like a country’s economy and decision support for intensive care units, as well as Special Relativistic effects (like time dilation) in 4-D will be presented.





    Intended Audience

    Attendees will acquire skills for the visualization of multidimensional problems and learn strategies for the discovery of patterns corresponding to RELATIONS in multivariate data. These patterns are much more compact and informative than looking at the data. They are robust in the presence of noise providing a Topology for proximity...


    Alfred Inselberg, Tel Aviv University
    Pei Ling Lai, Southern Taiwan University of Science and Technology

    Back to Top
    Go Back