Presentation / Installation
Uncertainty-Aware Multidimensional Scaling
DescriptionOur uncertainty-aware dimensionality reduction technique is an extension of multidimensional scaling for projecting sets of multivariate probability distributions. Uncertain data items in high-dimensional data sets are modeled as random vectors and then projected to low-dimensional space (e.g. 2D) for visualization. Uncertainty is propagated to obtain a mapping from high- to low-dimensioal distributions.
Contributor