arxivstat.MLcs.LG2026-06-29
Adjusted Wasserstein distances for bridging empirical and true distributions with applications to MDS
Flor Martinez-Sermeno, Arturo Jaramillo, Johan Van Horebeek
This paper examines how metric adjustments to Multidimensional Scaling (MDS) can enhance its effectiveness as a visual tool for pattern recognition. The distance under consideration, referred to as Max-D-SW, is an adjustment of the Max-Sliced Wasserstein distance. In contrast to…