Scholarlink

Proposes Frequency Adaptive Economic Forecasting (FAEF) to address non-stationarity challenges in economic time series prediction. Our innovations include a frequency decomposition strategy, adaptive frequency identification, and model-agnostic preprocessing that enables stage-aware economic forecasting. Experiments show 15%–43% accuracy gains and up to 42% fewer errors during crises. Frequency analysis validates its adaptability to non-stationary economic data.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

FAEF: Frequency-Adaptive Economics Forecasting that Works

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jiahui Zhang,
  • Zeqian Huang,
  • Jinxuan Wu,
  • Jiaxin Kang

摘要

Proposes Frequency Adaptive Economic Forecasting (FAEF) to address non-stationarity challenges in economic time series prediction. Our innovations include a frequency decomposition strategy, adaptive frequency identification, and model-agnostic preprocessing that enables stage-aware economic forecasting. Experiments show 15%–43% accuracy gains and up to 42% fewer errors during crises. Frequency analysis validates its adaptability to non-stationary economic data.