notesum.ai

Published at December 9

When Dimensionality Reduction Meets Graph (Drawing) Theory: Introducing a Common Framework, Challenges and Opportunities

cs.LG

Released Date: December 9, 2024

Authors: Fernando Paulovich, Alessio Arleo, Stef van den Elzen

Arxiv: http://arxiv.org/pdf/2412.06555v1

[Uncaptioned image]

Technique	Modeled Relationships
t-SNE	Joint probability of data items to be neighbors.
UMAP	Probability of data items to be neighbors.
ISOMAP	Geodesic distance given a nearest neighbor graph.
MDS	Pairwise high-dimensional distances.
Sammon’s Mapping	Pairwise high-dimensional distances.
LAMP	Pairwise weighted high-dimensional distance between data items in a sample and high-dimensional distance to these samples and other items in the dataset.
IDMAP	Pairwise high-dimensional distance.
LSP	Weighted high-dimensional distance between nearest neighbors.
ProjClus	Pairwise high-dimensional distances between data items in the same cluster
LLE	Weighted high-dimensional distance between nearest neighbors.
LoCH	Weighted high-dimensional distance between nearest neighbors.
Laplacian Eigenmaps	Pairwise distance between nearest neighbors.
PLSP	Weighted high-dimensional distance between nearest neighbors