Get Started
Research questionHow should distance be defined for high-dimensional clustering when Euclidean geometry mishandles anisotropy and feature interactions?In high-dimensional data, Euclidean distance can obscure meaningful relationships when feature variation is anisotropic or features interact. Redundant and constant features can further distort pairwise similarity and clustering behavior.
Machine Learning
Statistical Machine Learning
Latest papersRecent research connected to this question, newest first.Anisotropic View Distance Metric for High-Dimensional Data: Theory, Geometry, and Fast ComputationThe source studies a projection-based distance for K-Means-style similarity measurement, with full and selected two-dimensional projections. It provides metric and norm analyses, a matching-based computational reduction, and experiments across 12 datasets comparing the distance with Euclidean and other Lp metrics.research paper · Sep 3, 2026
Related questions
When can a random projection preserve JL distance bounds yet lose geometric variation needed for rankings and inference?How can clustering combine numerical and categorical variables while preserving an interpretable hierarchy?How can agnostic learning handle low-intrinsic-dimensional concepts under arbitrary distributions with Gaussian-robust benchmarks?How can image retrieval rankings capture neighborhood context in high-dimensional feature spaces?
Home
Topics
Search
Library