Robust OT-Guided Generative Residual Domain Adaptation for Bike-Sharing Demand Prediction under Temporal Domain Shift

2026-05-21

AI Summary

This study addresses the performance degradation of bike-sharing demand forecasting models caused by temporal shifts in mobility patterns. Framing March 2021–2026 Citi Bike demand prediction as a temporal domain adaptation task, the authors propose the Gen-ROTDA framework, which enhances robustness to noise and anomalies by transferring residual demand (rather than raw values) anchored at station-time points. The approach integrates robust optimal transport (ROT), generative residual modeling, and high-cost matching pruning to improve generalization under distributional shifts. Experimental results demonstrate that Gen-ROTDA achieves the lowest MAE on the primary 2025–2026 forecasting task and outperforms existing optimal transport–based methods on average across multiple tasks, with particularly pronounced gains when the target domain contains unlabeled anomalous data.

Abstract

Bike-sharing models trained on historical station-hour data may degrade when deployed in later years because travel patterns change over time. This paper studies March Citi Bike demand prediction from 2021 to 2026 as a temporal domain adaptation problem and proposes Gen-ROTDA, a robust optimal transport-guided residual domain adaptation framework. The method fits a target-domain station-time anchor with a small labeled target subset, transfers residual rather than raw demand, applies a deterministic label-preserving residual feature generator, and trims high-cost transport matches before training the final residual predictor. Experiments compare Gen-ROTDA with anchor-only, source-only, target-only, fine-tuning, MMD adaptation, Sinkhorn OTDA, ROTDA, and Gen-OTDA. Gen-ROTDA achieves the lowest MAE on the main 2025 to 2026 task and is the best OT-family method on average across multi-year tasks, although fine-tuning and MMD adaptation remain strong overall baselines. Under abnormal target-unlabeled records, Gen-ROTDA is much more stable than non-robust OT variants, suggesting that robust transport is useful for noisy temporal transfer in bike-sharing demand prediction.

Problem

• temporal domain shift
• bike-sharing demand prediction
• domain adaptation
• optimal transport
• time-series forecasting

Innovation

• optimal transport
• domain adaptation
• residual learning
• temporal shift
• robustness