[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Subscribe]
DM: PC plotsFrom: Warren Sarle Date: Thu, 7 Aug 1997 18:00:27 -0400 (EDT) Ronny Kohavi writes: > I met with him last year and posed the same question of how to > visualize 10,000 rules to a user. He did not have a simple answer > (nor do I expect one). > > As much as I find the area of parallel coordinates fascinating, it > takes users a long time to understand what is in the scene. That's > one reason why there's a 3-month course at Tel-Aviv univ on this > topic. There's beautiful theory behind it (e.g., how to pick the > order), but I don't see it becoming mainstream visualization for >data > mining/viz. IMHO, business users will not easily understand this > visualization. There have been many interesting developments in the theoretical aspects of parallel coordinate plots, and I can imagine that it might take a 3-month course to go into the subject in detail. But the basic idea of a parallel coordinate plot is very easy to understand. Researchers and practitioners in clinical psychology, education, and other fields have used simple parallel coordinate plots for decades, but they call these plots "profiles". To be more precise, a "profile" is a plot of a single case using parallel coordinates. Recent advances in computer graphics have made it possible to put dozens or perhaps hundreds of profiles in a single plot. A plot of hundreds of profiles can indeed be confusing. But I think even a "business user" could easily understand a plot showing 1 or 2 profiles, and then work up to plots with more numerous profiles. Displaying and understanding 10,000 profiles is, as Ronny suggests, quite a challenge. References: Greenhouse, S.W., and Geisser, S. (1959), "On methods in the analysis of profile data," Psychometrika, 24, 95-112. Hartigan, J.A. (1975), Clustering Algorithms, New York: John Wiley & Sons, Inc. -- Warren S. Sarle SAS Institute Inc. The opinions expressed here saswss@unx.sas.com SAS Campus Drive are mine and not necessarily (919) 677-8000 Cary, NC 27513, USA those of SAS Institute. * Do not send me unsolicited commercial, political, or religious email *
|
MHonArc
2.2.0