New MAPS metric quantifies multivariable time series predictability

Researchers have introduced the Multivariable Predictability Score (MAPS), a novel metric designed to quantify the intrinsic predictability of multivariable time series forecasting tasks. Unlike existing measures that may struggle with nonlinear dependencies or are limited to univariate series, MAPS utilizes covariate information to assess predictability across various lookback windows and forecasting horizons. This model-agnostic approach provides a quantitative measure of the predictive information available, and its effectiveness has been demonstrated through evaluations on synthetic and real-world datasets, showing close alignment with expected predictability. AI

IMPACT Provides a new tool for evaluating the inherent difficulty of time series forecasting tasks, potentially improving model development and comparison.

RANK_REASON The cluster contains a research paper detailing a new metric for time series analysis. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Gotit.pub
• Hugging Face
• IArxiv
• Multivariable Predictability Score
• ScienceCast

• paper
• other

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