Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Positional-aware Spatio-Temporal Network for Large-Scale Traffic Prediction

Runfei Chen

TL;DR

A lightweight Positional-aware Spatio-Temporal Network (PASTN) is proposed to effectively capture both temporal and spatial complexities in an end-to-end manner and introduces positional-aware embeddings to separate each node's representation.

Abstract

Traffic flow forecasting has emerged as an indispensable mission for daily life, which is required to utilize the spatiotemporal relationship between each location within a time period under a graph structure to predict future flow. However, the large travel demand for broader geographical areas and longer time spans requires models to distinguish each node clearly and possess a holistic view of the history, which has been paid less attention to in prior works. Furthermore, increasing sizes of data hinder the deployment of most models in real application environments. To this end, in this paper, we propose a lightweight Positional-aware Spatio-Temporal Network (PASTN) to effectively capture both temporal and spatial complexities in an end-to-end manner. PASTN introduces positional-aware embeddings to separate each node's representation, while also utilizing a temporal attention module to improve the long-range perception of current models. Extensive experiments verify the effectiveness and efficiency of PASTN across datasets of various scales (county, megalopolis and state). Further analysis demonstrates the efficacy of newly introduced modules either.

Positional-aware Spatio-Temporal Network for Large-Scale Traffic Prediction

TL;DR

A lightweight Positional-aware Spatio-Temporal Network (PASTN) is proposed to effectively capture both temporal and spatial complexities in an end-to-end manner and introduces positional-aware embeddings to separate each node's representation.

Abstract

Traffic flow forecasting has emerged as an indispensable mission for daily life, which is required to utilize the spatiotemporal relationship between each location within a time period under a graph structure to predict future flow. However, the large travel demand for broader geographical areas and longer time spans requires models to distinguish each node clearly and possess a holistic view of the history, which has been paid less attention to in prior works. Furthermore, increasing sizes of data hinder the deployment of most models in real application environments. To this end, in this paper, we propose a lightweight Positional-aware Spatio-Temporal Network (PASTN) to effectively capture both temporal and spatial complexities in an end-to-end manner. PASTN introduces positional-aware embeddings to separate each node's representation, while also utilizing a temporal attention module to improve the long-range perception of current models. Extensive experiments verify the effectiveness and efficiency of PASTN across datasets of various scales (county, megalopolis and state). Further analysis demonstrates the efficacy of newly introduced modules either.
Paper Structure (20 sections, 12 equations, 6 figures, 6 tables)

This paper contains 20 sections, 12 equations, 6 figures, 6 tables.

Table of Contents

  1. Abstract
  2. 1 Introduction
  3. 2 Related Work
  4. 3 Preliminaries
  5. 4 Methodology
  6. 4.1 Spatial-Positional-Aware Embedding
  7. 4.2 Spatiotemporal Learning Module
  8. 4.3 Temporal-Positional-Aware Module
  9. 4.4 Loss Function
  10. 5 Experiments
  11. 5.1 Experiment Settings
  12. 5.2 Effectiveness
  13. 5.3 Generalization
  14. 5.4 Ablations
  15. 5.5 Case Study and Visualization
  16. ...and 5 more sections

Figures (6)

  • Figure 1: Detailed framework of PASTN.
  • Figure 2: Performance vs. inference time on GLA and CA datasets. The bubble size represents the learnable parameter volume.
  • Figure 3: MAE, RMSE and MAPE on SD-19 dataset.
  • Figure 4: Embedding visualization of SPAE. A more uniform and complete circular distribution of colors indicates lower similarity among nodes.
  • Figure 5: Attention matrix (Left) and output embedding visualization of TPAM (Right).
  • ...and 1 more figures