Title: Section pursuit
Authors: Ursula Laa - BOKU University (Austria) [presenting]
Di Cook - Monash University (Australia)
Andreas Buja - University of Pennsylvania (United States)
German Valencia - Monash (Australia)
Abstract: Multivariate data is often visualized using linear projections, produced by techniques such as principal component analysis, linear discriminant analysis, and projection pursuit. A problem with projections is that they obscure low and high-density regions near the center of the distribution. Sections, or slices, can help to reveal them. Section pursuit (building on the extensive work in projection pursuit) is introduced, a new method to search for interesting slices of the data. Linear projections are used to define sections of the parameter space, and we calculate interestingness by comparing the distribution of observations, inside and outside a section. By optimizing this index, it is possible to reveal features such as holes (low density) or grains (high density), which can be useful when data distributions depart from uniform or normal, as in visually exploring nonlinear manifolds, and functions in multivariate space. We will show how section pursuit can be applied when exploring decision boundaries from classification models or when exploring subspaces induced by complex inequality conditions from a multiple parameter model.