Shape-Preserving Dimensionality Reduction : An Algorithm and Measures of Topological Equivalence

We introduce a linear dimensionality reduction technique preserving topological features via persistent homology. The method is designed to find linear projection which preserves the persistent diagram of a point cloud via simulated annealing. The projection induces a set of canonical simplicial maps from the Rips (or \v{C}ech) filtration of to that of . In addition to the distance between persistent diagrams, the projection induces a map between filtrations, called filtration homomorphism. Using the filtration homomorphism, one can measure the difference between shapes of two filtrations directly comparing simplicial complexes with respect to quasi-isomorphism or strong homotopy equivalence . These and measures how much portion of corresponding simplicial complexes is quasi-isomorphic or homotopy equivalence respectively. We validate the effectiveness of our framework with simple examples.
View on arXiv