Forecasting traffic flows in irregular regions with multi-graph convolutional network and gated recurrent unit

Seng, Dewen; Lv, Fanshun; Liang, Ziyi; Shi, Xiaoying; Fang, Qiming

doi:10.1631/fitee.2000243

Cited by 23 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The advent of deep learning has led to the emergence of various methods, such as the Recurrent Neural Network (RNN) and convolutional neural network (CNN) based approaches. RNN-based methods commonly utilize specialized architectures like Long Short-Term Memory (LSTM) or Gated Recurrent Units (GRUs) for effective modeling and feature extraction [7,8,9]. The convolutional neural network method converts the trajectory data into image form, uses a convolutional neural network to extract spatialtemporal features [10], and then fuses and classifies the features through the fully connected layer.…”

Section: Related Workmentioning

confidence: 99%

Spatiotemporal Features Recognition and Prediction of Urban Hot Spots Based on Trajectory Data

Cao,

Sun

2024

Advances in Transdisciplinary Engineering

View full text Add to dashboard Cite

With the continuous progress of urbanization, the construction of cities is gradually accelerated, and the urban population density is constantly increasing. Hot spots represent regions with more trips, greater population flow and higher travel demand. With the increasing complexity of traffic data, traditional methods for predicting hot spots can no longer meet the current situation. This also provides opportunities for the development of deep learning techniques. Given the continuous expansion in scale and intricacy of trajectory data, its relevance has become increasingly significant in domains such as urban planning, traffic management, and business decision-making. Aiming at the hot spot identification problem of trajectory data, this paper, inspired by the traffic prediction model, applied the STGCN model to the urban hot spot prediction problem and improved it. The model consists of four parts: cluster processing, spatial feature extraction, time feature extraction and hot spot region prediction. The experimental outcomes demonstrate that the model exhibits superior predictive performance across diverse datasets compared to conventional methods, showcasing robust generalization capabilities and practical applicability.

show abstract

Section: Related Workmentioning

confidence: 99%

Spatiotemporal Features Recognition and Prediction of Urban Hot Spots Based on Trajectory Data

Cao,

Sun

2024

Advances in Transdisciplinary Engineering

View full text Add to dashboard Cite

show abstract

“…(1) Historical Average Model (HA) [5] (2) Auto-Regressive Integrated Moving Average (ARIMA) (3) Time Graph Convolution (T-GCN) [20] (4) Convolutional Long and Short Term Memory (Conv-LSTM) [21] (5) ST-ResNet [22] (6) DeepSD [23] (7) Improved VMD-GAT(5)-BiGRU (ours) As shown in Table 6, the HA and ARIMA models do not extract the spatial features among nodes but focus on mining the temporal features in the traffic flow data. T-GCN and ConvLSTM consider the spatiotemporal correlation of traffic flow data simultaneously but do not guide the neural network with a priori features.…”

Section: E Baseline Experimentsmentioning

confidence: 99%

“…A framework for Spatio-temporal correlation capture was thus constructed, which has set off a boom in traffic flow prediction based on Spatio-temporal correlation. Seng et al [21] used multigraph convolutional networks and Gated Recurrent Unit (GRU) networks to capture the Spatiotemporal correlation of unstable regions in road networks. Wan et al [22] extracted the Spatio-temporal features of rule-based raster data using 3D-CNN and recalibration blocks, then verified the model's performance under traffic prediction tasks using real data sets.…”

Section: Introductionmentioning

confidence: 99%

ST-VGBiGRU: A Hybrid Model for Traffic Flow Prediction With Spatio-Temporal Multimodality

Yin

Liu

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Traffic flow prediction is a key step to realizing the effective guidance and control of intelligent transportation systems. For using the short-term non-stationarity and spatio-temporal correlation presented in traffic flow, a spatio-temporal hybrid prediction model, which called ST-VGBiGRU, based on improved Variational Modal Decomposition (VMD), Graph Attention Network (GAT), and Bidirectional Gated Recurrent Unit (BiGRU) network is proposed. First, the traffic flow sequence is decomposed into a series of relatively stationary modal components using VMD algorithm to reduce its short-term nonstationarity. The high-frequency modal components are noise-reduced using the Fuzzy Entropy (Fuzzy En) method to improve the accuracy of decomposition. After that, the GAT network is used to capture the different attention levels of the prediction node to their neighboring traffic nodes, which obtain more spatial characteristics of traffic flow. Then, each modal component containing spatial features is fed into the BiGRU network separately to capture its temporal correlation. Each model parameter is trained to the optimum using the improved RMSProp algorithm, which improves the model's prediction accuracy while speeding up the convergence of RMSProp algorithm. In order to illustrate the performance of the ST-VGBiGRU model, the RTMC traffic dataset is used to conduct ablation experiments on the improved VMD module, the GAT module, and the improved BiGRU module. Meanwhile, we combined the PeMS traffic dataset to conduct baseline experiments and multi-step prediction experiments with the other six models. The results show that the prediction performance of our model is better than all the other baseline models.

show abstract

“…Fusco et al proposed short-term traffic predictions on large urban traffic networks by using applications of network-based machine learning models and dynamic traffic assignment models [15]. Seng [16] et al used a deep neural network and a regular grid cyclic neural network to capture the spatial dependence of traffic flow prediction and proposed an irregular regional traffic flow prediction model based on a multi-graph convolution network and gated cyclic unit (MGCN-GRU). Hashemi et al proposed real-time traffic network state estimation and prediction with decision support capabilities for applications to integrated corridor management [17].…”

Section: Introductionmentioning

confidence: 99%

Prediction of Highway Blocking Loss Based on Ensemble Learning Fusion Model

Guo

Zhang

et al. 2022

Electronics

View full text Add to dashboard Cite

Road blocking events refer to road traffic blocking caused by landslides, debris flow, snow disasters, rolling stones and other factors. To predict road blocking events, the limit gradient lifting model (XGBoost), random forest regression model (RF regression) and support-vector regression model (SVR) are used as the prediction meta-models, and then the meta-models are fused by a logical regression algorithm to construct a road blocking loss prediction fusion model based on ensemble learning. The actual road blocking event data are used to train the model. Using the same blocking location and similar blocking loss characteristics between adjacent points to fill in the missing value and conducting one-hot encoding for other short character sets with obvious category characteristics such as letters, numbers, and Chinese characters overcomes the problems of inherent data loss, error and time logic disorder in the blocking event data set. The test results show that the R2 score based on the stacking fusion model reaches 0.91, which is 18% higher than RF and 11% and 5.8% higher than SVR and XGBoost, respectively, and the RMSE and MAE values are 0.1707 and 0.0341, respectively. Therefore, the proposed road blocking data preprocessing method and road blocking loss prediction fusion model can be used to predict the amount of blocking event loss.

show abstract

Forecasting traffic flows in irregular regions with multi-graph convolutional network and gated recurrent unit

Cited by 23 publications

References 20 publications

Spatiotemporal Features Recognition and Prediction of Urban Hot Spots Based on Trajectory Data

Spatiotemporal Features Recognition and Prediction of Urban Hot Spots Based on Trajectory Data

ST-VGBiGRU: A Hybrid Model for Traffic Flow Prediction With Spatio-Temporal Multimodality

Prediction of Highway Blocking Loss Based on Ensemble Learning Fusion Model

Contact Info

Product

Resources

About