STLGRU: Spatio-Temporal Lightweight Graph GRU for Traffic Flow Prediction

TL;DR

This work proposes Spatio-Temporal Lightweight Graph GRU, namely STLGRU, a novel traffic forecasting model for predicting traffic flow accurately that can effectively capture dynamic local and global spatial-temporal relations of traffic networks using memory-augmented attention and gating mechanisms in a continuously synchronized manner.

Abstract

Reliable forecasting of traffic flow requires efficient modeling of traffic data. Indeed, different correlations and influences arise in a dynamic traffic network, making modeling a complicated task. Existing literature has proposed many different methods to capture traffic networks' complex underlying spatial-temporal relations. However, given the heterogeneity of traffic data, consistently capturing both spatial and temporal dependencies presents a significant challenge. Also, as more and more sophisticated methods are being proposed, models are increasingly becoming memory-heavy and, thus, unsuitable for low-powered devices. To this end, we propose Spatio-Temporal Lightweight Graph GRU, namely STLGRU, a novel traffic forecasting model for predicting traffic flow accurately. Specifically, our proposed STLGRU can effectively capture dynamic local and global spatial-temporal relations of traffic networks using memory-augmented attention and gating mechanisms in a continuously synchronized manner. Moreover, instead of employing separate temporal and spatial components, we show that our memory module and gated unit can successfully learn the spatial-temporal dependencies with reduced memory usage and fewer parameters. Extensive experimental results on three real-world public traffic datasets demonstrate that our method can not only achieve state-of-the-art performance but also exhibit competitive computational efficiency. Our code is available at https://github.com/Kishor-Bhaumik/STLGRU

Figures (5)

• Figure 1: Overall architecture of STLGRU designed for multivariate traffic forecasting. Our model consists of a memory-augmented attention module and a gated unit, which capture the long-range local and global dependencies. It takes input from a single time step with an initial hidden state and outputs a hidden state for the next time step.
• Figure 2: Graph generation from learnable node embeddings $E$
• Figure 3: Performance comparison of spatio-temporal models and STLGRU with different settings. MAE, RMSE and MAPE of 1-hour forecasting on three datasets are plotted.
• Figure 4: Visualization of the predicted traffic flow.
• Figure 5: Sensor Distribution of three traffic datasets, where the dots are the traffic-sensor locations.