Piecewise rectangular visualization maps: optimizing areas and similarities between portions
In this talk, we propose the use of MINLP to build visualization maps for a set of individuals which have attached similarities as well a statistical value. In these maps, each individual is represented as a connected portion made of pixels on a grid. The map is built in such a way that distances between portions resemble the similarities between the individuals and the areas of the portions are similar to the statistical value attached to the individuals.
Keywords: Visualization heuristics mixed integer nonlinear programming