PRADA: Prompt-guided Representation Alignment and Dynamic Adaption for time series forecasting

Yinhao Liu, Zhenyu Kuang, Hongyang Zhang, Chen Li, Feifei Li, Xinghao Ding

Abstract: Highlights•A novel method named PRADA is proposed for time series forecasting.•The TSAA module performs semantic alignment between time series and text using learnable prompts.•Time-Frequency Dual Constraint is designed to capture the label autocorrelation of time series.

External IDs:doi:10.1016/j.knosys.2025.113478