Deep spatial-temporal networks for crowd flows prediction by dilated convolutions and region-shifting attention mechanis
- PDF / 1,281,969 Bytes
- 14 Pages / 595.224 x 790.955 pts Page_size
- 101 Downloads / 204 Views
Deep spatial-temporal networks for crowd flows prediction by dilated convolutions and region-shifting attention mechanism Chujie Tian1 · Xinning Zhu2 · Zheng Hu1 · Jian Ma1
© Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Flow prediction at a citywide level is of great significance to traffic management and public safety. Since deep learning has achieved success to deal with complex nonlinear problems, it has drawn increasing attention on making crowd flows prediction through neural networks. Generally, convolutional neural network (CNN) and recurrent neural network (RNN) have been applied to model the spatial-temporal dependency of the city. However, there are still two major challenges in predicting flows. First, it is difficult to train the model with the ability to capture both the nearby and distant spatial dependency by deep local convolutions. Second, daily and weekly patterns in temporal dependency are not strictly periodic for their dynamic temporal shifting in each region. To address these issues, we propose a novel deep learning model which called Local-Dilated Region-Shifting Network (LDRSN). LDRSN combines local convolutions with dilated convolutions to learn the nearby and distant spatial dependency. Furthermore, a new region-level attention mechanism is proposed to model the temporal shifting which varies by region. In the experiments, we compare the proposed method with other stateof-the-art methods in two real-world crowd flows datasets. The Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE) were used as the evaluation indexes. The experiment results show the effectiveness of the proposed model. Keywords Dilated convolution · Region-level attention mechanism · Crowd flows prediction · Spatial-temporal network
1 Introduction Traffic prediction has been regarded as the foundation of intelligent traffic management and transport policy [1]. The accurate and timely prediction of traffic information plays a crucial role for many levels of intelligent transportation system (ITS) [2]. The prediction information allows individuals make better travel plans and improves the travel experience by reducing the cost of money and time [3]. It also benefits transportation management by revealing problematic locations, and improves the efficiency and
Zheng Hu
[email protected] 1
Stat Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing, China
2
Key Laboratory of Universal Wireless Communications, Ministry of Education, Beijing University of Posts and Telecommunications, Beijing, China
safety of the road by the scientific decision-making [4]. Furthermore, transport policies such as tourism prices and vehicle controls are evaluated and revised by traffic information prediction. As traffic information prediction is an important issue for ITS, it has generated great interests among researchers. Regression methods have frequently been used in past decades [5]. Autoregressive
Data Loading...
