Spatiotemporal Prediction of Secondary Crashes by Rebalancing Dynamic and Static Data with Generative Adversarial Networks

2025-01-17

AI Summary

Addressing the challenges of extreme class imbalance, complex coupling between dynamic and static features, and variable-length temporal sequences in secondary traffic accident prediction, this paper proposes VarFusiGAN-Transformer—a novel end-to-end joint prediction framework. The architecture innovatively integrates LSTM with generative adversarial networks (GANs) to jointly model and adversarially generate dynamic time-series and static features. To enhance representation learning for minority classes, it introduces auxiliary discriminators and a dedicated static feature generator. A Transformer-based prediction head unifies probabilistic accident occurrence forecasting with precise spatiotemporal localization. Evaluated on a real-world traffic dataset, the method achieves an 18.7% improvement in accuracy and a 22.3% gain in F1-score over state-of-the-art imbalanced learning and spatiotemporal forecasting approaches, demonstrating superior performance in handling data scarcity and heterogeneous feature integration.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Generative Adversarial Networks (GANs) for Vision

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Data imbalance is a common issue in analyzing and predicting sudden traffic events. Secondary crashes constitute only a small proportion of all crashes. These secondary crashes, triggered by primary crashes, significantly exacerbate traffic congestion and increase the severity of incidents. However, the severe imbalance of secondary crash data poses significant challenges for prediction models, affecting their generalization ability and prediction accuracy. Existing methods fail to fully address the complexity of traffic crash data, particularly the coexistence of dynamic and static features, and often struggle to effectively handle data samples of varying lengths. Furthermore, most current studies predict the occurrence probability and spatiotemporal distribution of secondary crashes separately, lacking an integrated solution. To address these challenges, this study proposes a hybrid model named VarFusiGAN-Transformer, aimed at improving the fidelity of secondary crash data generation and jointly predicting the occurrence and spatiotemporal distribution of secondary crashes. The VarFusiGAN-Transformer model employs Long Short-Term Memory (LSTM) networks to enhance the generation of multivariate long-time series data, incorporating a static data generator and an auxiliary discriminator to model the joint distribution of dynamic and static features. In addition, the model's prediction module achieves simultaneous prediction of both the occurrence and spatiotemporal distribution of secondary crashes. Compared to existing methods, the proposed model demonstrates superior performance in generating high-fidelity data and improving prediction accuracy.

Problem

• Secondary Accident Prediction
• Data Imbalance
• Dynamic and Static Information Integration

Innovation

• VarFusiGAN-Transformer
• GAN-Transformer Integration
• Dynamic-Static Information Processing