Open AccessGraph Wavelet Long Short-Term Memory Neural Network: A Novel Spatial-Temporal Network for Traffic Prediction
Author(s) -
Qiong Wu,
Qin Fu,
Mingxin Nie
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1549/4/042070
Subject(s) - computer science , graph , long short term memory , wavelet , artificial neural network , artificial intelligence , term (time) , deep learning , recurrent neural network , data mining , theoretical computer science , physics , quantum mechanics
Timely accurate traffic prediction is important in the Intelligent Traffic System (ITS). It has time-varying traffic patterns and the complicated spatial dependencies on traffic network topology which makes the prediction challenging. In this paper, we propose a novel deep learning framework—Graph Wavelet Long Short-Term Memory Neural Network (GWNN-LSTM) to capture the spatial and temporal dependence simultaneously. Moreover, Graph Wavelet Neural Network (GWNN) is utilized for spatial correlations and Long Short-Term Memory Neural Network (LSTM) is used to capture the dynamic temporal correlations in traffic time series data. Experiments on real-world datasets from loop detectors in the highway of Los Angeles County (METR-LA) demonstrate that the proposed GWNN-LSTM model outperforms the state-of-the-art baselines.