A Spectral Method for Assessing and Combining Multiple Data Visualizations. (arXiv:2210.13711v1 [stat.ML])

Dimension reduction and data visualization aim to project a high-dimensional
dataset to a low-dimensional space while capturing the intrinsic structures in
the data. It is an indispensable part of modern data science, and many
dimensional reduction and visualization algorithms have been developed.
However, different algorithms have their own strengths and weaknesses, making
it critically important to evaluate their relative performance for a given
dataset, and to leverage and combine their individual strengths. In this paper,
we propose an efficient spectral …

arxiv data data visualizations