Using ensemble learning with hybrid graph neural networks and transformers to predict traffic in cities

TL;DR

HybridST combines graph neural networks, transformers, and ensemble learning to improve city traffic prediction accuracy across diverse datasets, supporting real-time urban mobility planning.

Contribution

This paper introduces HybridST, a novel hybrid architecture integrating GNNs, transformers, and ensemble methods for enhanced traffic prediction in complex urban environments.

Findings

HybridST outperforms classical baselines on key metrics.

Model is scalable and easy to interpret.

Effective across multiple public benchmark datasets.

Abstract

Intelligent transportation systems (ITS) still have a hard time accurately predicting traffic in cities, especially in big, multimodal settings with complicated spatiotemporal dynamics. This paper presents HybridST, a hybrid architecture that integrates Graph Neural Networks (GNNs), multi-head temporal Transformers, and supervised ensemble learning methods (XGBoost or Random Forest) to collectively capture spatial dependencies, long-range temporal patterns, and exogenous signals, including weather, calendar, or control states. We test our model on the METR-LA, PEMS-BAY, and Seattle Loop tree public benchmark datasets. These datasets include situations ranging from freeway sensor networks to vehicle-infrastructure cooperative perception. Experimental results show that HybridST consistently beats classical baselines (LSTM, GCN, DCRNN, PDFormer) on important metrics like MAE and RMSE, while still being very scalable and easy to understand. The proposed framework presents a promising avenue for real-time urban mobility planning, energy optimization, and congestion alleviation strategies, especially within the framework of smart cities and significant events such as the 2030 FIFA World Cup.