Reproduce “High-Dimensional Multivariate Forecasting with Low-Rank Gaussian Copula Processes” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

David Salinas, Michael Bohlke-Schneider, Laurent Callot, Roberto Medico, Jan Gasthaus

GP-Copula produces joint probabilistic forecasts for many related time series by updating a recurrent state for each series and using those states to parameterize a time-varying, low-rank Gaussian-copula covariance. Across six datasets, the authors report average CRPS and CRPS-Sum improvements of 10% and 40% over the second-best methods. These are averages, not wins on every dataset; Taxi is an exception.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.