Reproduce “TIGER: Time-Series Classification with In-Context-Learning Gated Ensemble of Representations” 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
Johann Faouzi
TIGER classifies univariate time series by extracting four feature representations and applying the same three classifiers to each, then combining their predictions with a rule selected by training samples per class. On 142 UCR data sets averaged over 30 resamples, TIGER achieved mean accuracy 0.884 versus 0.876 for HIVE-COTE 2.0; this comparison does not establish performance on multivariate series.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.