WEASEL 2.0 time series classifier revisited with adaptive ensemble rule

Researchers have revisited the WEASEL 2.0 time series classification algorithm, confirming its performance on 114 UCR datasets with a mean accuracy of 0.865. The study found that while the downstream classifier, feature weighting, and maximum window size rules were robust, the maximum ensemble size rule was over-provisioned for longer series. This led to the development of an adaptive rule that adjusts ensemble size based on series length and class count, significantly reducing memory and computation time with minimal impact on accuracy. AI

IMPACT This research offers a more efficient approach to time series classification, potentially reducing computational costs for related AI tasks.

RANK_REASON The item is a research paper detailing an improvement to an existing algorithm. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXivLabs
• CatalyzeX Code Finder for Papers
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• IArxiv Recommender
• Influence Flower
• ScienceCast
• UCR datasets
• WEASEL 2.0
• Wilcoxon signed-rank test

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →