Optimal copula transport: clustering multivariate time series via Wasserstein distance
FrnkNlsn · x · 2026-08-21
A recommendation of the ICASSP 2016 paper "Optimal Copula Transport for Clustering Multivariate Time Series" (arXiv:1509.08144, by Gautier Marti, Frank Nielsen, Philippe Donnat). Copulas fully capture the dependence of multivariate distributions by factorizing the joint density with 1D marginals. The paper uses optimal transport (Wasserstein distance) between copulas to encode both intra-dependence of a multivariate time series and inter-dependence between two series, defining two distances for clustering. The proposed multivariate dependence coefficient is robust to noise, deterministic, and can target specified dependencies.
More from Research
- Hugging Face discusses training models to replicate AI research papers — Hugging Face · 2026-08-21
- NeurIPS 2026 Sydney launches call for papers on robot learning with world models — PontiEdoardo · 2026-08-21
- Cohere Session: From Face ID to Cat Detection in Computer Vision — Cohere_Labs · 2026-08-21
- Smokebench: Lightweight TUI for quick local LLM benchmarking — Ninja-5000 · 2026-08-21
- Metasurface and neural network combo enables compact multi-dimensional infrared imaging — bravo_abad · 2026-08-21
- The Enduring Value of Math in an Age of AI — stevenstrogatz · 2026-08-21