Abstract
AbstractThe paper deals with the multidimensional scaling (MDS) that depends on the class of nonlinear projection methods for a visual representation of multidimensional data. The performance of a new MDS-type method for multidimensional data dimensionality reduction and visualization (Geometric MDS) has been investigated visually using GeoGebra. Dynamic geometry program GeoGebra is a non-commercial and interactive software for the visual representation of algebra and geometry. We made specific GeoGebra scripts for the visual representation of the convergence process of Geometric MDS. This allows us to analyze the optimization of the stress function visually, describing the visualization quality and find the basins of attraction to the local minima. The results allow an easier comprehension of the MDS stress optimization by the anti-gradient search. Moreover, the results deepen the understanding of Geometric MDS, in general.KeywordsVisual analysisMultidimensional scalingGeoGebraDimensionality reductionGeometric MDS
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
