Bias-Corrected Adaptive Conformal Inference for Multi-Horizon Time Series Forecasting
Ankit Lade, Sai Krishna J., Indar Kumar
TLDR
Bias-Corrected Adaptive Conformal Inference (BC-ACI) improves time series prediction intervals by correcting for persistent forecast bias, leading to narrower, more accurate bands.
Key contributions
- Proposes Bias-Corrected ACI (BC-ACI) to address persistent forecast bias in time series prediction.
- Augments ACI with an online exponentially weighted moving average (EWM) to estimate and correct forecast bias.
- Re-centers prediction intervals and uses an adaptive dead-zone threshold to prevent degradation on well-calibrated data.
- Reduces Winkler interval scores by 13-17% under distribution shifts, maintaining performance on stationary data.
Why it matters
Existing Adaptive Conformal Inference (ACI) widens intervals when forecasts are biased, leading to overly conservative predictions. BC-ACI offers a more efficient and accurate way to handle bias, providing tighter and more reliable prediction intervals for time series data under distribution shifts. This improves the utility of forecasts in dynamic environments.
Original Abstract
Adaptive Conformal Inference (ACI) provides distribution-free prediction intervals with asymptotic coverage guarantees for time series under distribution shift. However, ACI only adapts the quantile threshold -- it cannot shift the interval center. When a base forecaster develops persistent bias after a regime change, ACI compensates by widening intervals symmetrically, producing unnecessarily conservative bands. We propose Bias-Corrected ACI (BC-ACI), which augments standard ACI with an online exponentially weighted moving average (EWM) estimate of forecast bias. BC-ACI corrects nonconformity scores before quantile computation and re-centers prediction intervals, addressing the root cause of miscalibration rather than its symptom. An adaptive dead-zone threshold suppresses corrections when estimated bias is indistinguishable from noise, ensuring no degradation on well-calibrated data. In controlled experiments across 688 runs spanning two base models, four synthetic regimes, and three real datasets, BC-ACI reduces Winkler interval scores by 13--17% under mean and compound distribution shifts (Wilcoxon p < 0.001) while maintaining equivalent performance on stationary data (ratio 1.002x). We provide finite-sample analysis showing that coverage guarantees degrade gracefully with bias estimation error.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.