Fuzzy simplicial sets and their application to geometric data analysis (2406.11154v1)

Abstract: In this article, we expand upon the concepts introduced in \cite{Spivak09} about the relationship between the category $\mathbf{UM}$ of uber metric spaces and the category $\mathbf{sFuz}$ of fuzzy simplicial sets. We show that fuzzy simplicial sets can be regarded as natural combinatorial generalizations of metric relations. Furthermore, we take inspiration from UMAP to apply the theory to dimension reduction (manifold learning) and data visualization, while refining some of their constructions to put the corresponding theory on a more solid footing. A generalization of the adjunction between $\mathbf{UM}$ and $\mathbf{sFuz}$ will allow us to view the adjunctions used in both publications as special cases. Moreover, we derive an explicit description of colimits in $\mathbf{UM}$ and the realization functor $\text{Re}:\mathbf{sFuz}\to\mathbf{UM}$, as well as rigorous definitions of functors that make it possible to recursively merge sets of fuzzy simplicial sets. We show that $\mathbf{UM}$ and the category of extended-pseudo metric spaces $\mathbf{EPMet}$ can be embedded into $\mathbf{sFuz}$ and provide a description of the adjunctions between the category of truncated fuzzy simplicial sets and $\mathbf{sFuz}$, which we relate to persistent homology. Combining those constructions, we can show a surprising connection between the well-known dimension reduction methods UMAP and Isomap and derive an alternative algorithm, which we call IsUMap, that combines some of the strengths of both methods. We compare it with UMAP and Isomap and provide explanations for observed differences.