FairTP: A Prolonged Fairness Framework for Traffic Prediction

TL;DR

FairTP addresses the lack of long-term fairness in traffic prediction by introducing two dynamic fairness notions, region-based static fairness ($RSF$) and sensor-based dynamic fairness ($SDF$). It couples a state-aware sampling mechanism with a flexible spatio-temporal predictor to reweight training data toward underrepresented regions and protract fairness over time. The framework optimizes a joint objective that balances predictive accuracy with short-term regional equity and long-term sensor consistency, demonstrated across two real-world datasets and multiple baselines. Results show substantial improvements in fairness metrics with limited or negligible losses in MAE, highlighting the practical potential for fairer, more equitable traffic management. The approach is modular and readily deployable alongside existing traffic forecasting models, offering tangible benefits for urban equity and resource planning.

Abstract

Traffic prediction plays a crucial role in intelligent transportation systems. Existing approaches primarily focus on improving overall accuracy, often neglecting a critical issue: whether predictive models lead to biased decisions by transportation authorities. In practice, the uneven deployment of traffic sensors across urban areas results in imbalanced data, causing prediction models to perform poorly in certain regions and leading to unfair decision-making. This imbalance ultimately harms the equity and quality of life for residents. Moreover, current fairness-aware machine learning models only ensure fairness at specific time points, failing to maintain fairness over extended periods. As traffic conditions change, such static fairness approaches become ineffective. To address this gap, we propose FairTP, a framework for prolonged fair traffic prediction. We introduce two new fairness definitions tailored for dynamic traffic scenarios. Fairness in traffic prediction is not static; it varies over time and across regions. Each sensor or urban area can alternate between two states: "sacrifice" (low prediction accuracy) and "benefit" (high prediction accuracy). Prolonged fairness is achieved when the overall states of sensors remain similar over a given period. We define two types of fairness: region-based static fairness and sensor-based dynamic fairness. To implement this, FairTP incorporates a state identification module to classify sensors' states as either "sacrifice" or "benefit," enabling prolonged fairness-aware predictions. Additionally, we introduce a state-guided balanced sampling strategy to further enhance fairness, addressing performance disparities among regions with uneven sensor distributions. Extensive experiments on two real-world datasets demonstrate that FairTP significantly improves prediction fairness while minimizing accuracy degradation.

Figures (7)

• Figure 1: Illustration of fairness issues in traffic prediction.
• Figure 2: The framework of the proposed FairTP.
• Figure 3: Regional performance on two datasets
• Figure 4: Effect of dynamic time length $T_d$.
• Figure 5: Effect of sampled number $N_{sam}$.