Reproduce “HGTS-Former: Hierarchical HyperGraph Transformer for Multivariate Time Series Analysis” 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
Hao Si, Xiao Wang, Fan Zhang, Xiaoya Zhou, Dengdi Sun, Wanli Lyu, Qingquan Yang, Jin Tang
HGTS-Former models multivariate time series by grouping temporal patterns within each variable, then learning higher-order relationships across variables with sparse attention. On EAST-ELM640, it reports higher F1 (79.7) and precision (80.8) than CrossLinear, but lower accuracy and AUC. Across forecasting tasks, it records 15 first-place MSE and 27 MAE results among 32 tasks, with exceptions including ETTm2.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.