Persistent Structure Meets Dynamic Attention: Cross-Variable Priors for Multivariate Time Series Forecasting
A Params-Per-Pair diagnostic is introduced that predicts from dataset properties alone whether structural priors will help and reveals a horizon-dependent complementarity: the structural prior contributes 33% of the gain at short horizons but 88% at long horizons, confirming that time-invariant knowledge compensates as temporal signal fades.
C. Mohapatra, Rohit Malshe, J. Pachón
· 0 citations