Article
Computer Science, Artificial Intelligence
Ge Zheng, Wei Koong Chai, Jing-Lin Duanmu, Vasilis Katos
Summary: This paper evaluates recent hybrid deep learning models for traffic prediction. The models are categorized based on their feature extraction methods and analyzed for their modules and designs. A performance comparison study is conducted on ten representative models, and findings show differences in prediction accuracy based on design decisions.
INFORMATION FUSION
(2023)
Article
Computer Science, Artificial Intelligence
Qiuyang Huang, Yongjian Yang, Yuanbo Xu, Funing Yang, Zhilu Yuan, Yongxiong Sun
Summary: Mobile signaling data has great value for urban traffic monitoring, improving coverage and accuracy.
Article
Green & Sustainable Science & Technology
Navin Ranjan, Sovit Bhandari, Pervez Khan, Youn-Sik Hong, Hoon Kim
Summary: This paper presents an inexpensive and efficient Traffic Congestion Pattern Analysis algorithm based on Image Processing, as well as a deep neural network architecture formed from Convolutional Autoencoder, both of which show high efficiency in predicting urban traffic conditions based on a case study conducted in Seoul city.
Article
Computer Science, Information Systems
Fangyuan Sun, Jia Yu, Rong Hao, Ming Yang, Fanyu Kong
Summary: The road network is crucial for guiding daily travel, and the shortest distance query is a fundamental operation. However, existing privacy-preserving schemes for large-scale encrypted graphs are designed for static graphs and cannot handle practical updates in road networks. To address this, we propose a dynamic incremental update algorithm for the encrypted graph that achieves adaptive semantic security and practicality.
IEEE SYSTEMS JOURNAL
(2023)
Article
Chemistry, Multidisciplinary
Sachin Ranjan, Yeong-Chan Kim, Navin Ranjan, Sovit Bhandari, Hoon Kim
Summary: Traffic congestion is a significant problem worldwide, and a hybrid deep neural network algorithm based on HRNet and ConvLSTM has been proposed for traffic congestion prediction. The model outperforms four other state-of-the-art architectures in terms of accuracy, precision, and recall.
APPLIED SCIENCES-BASEL
(2023)
Article
Engineering, Civil
Hyun Su Park, Yong Woo Park, Oh Hoon Kwon, Shin Hyoung Park
Summary: This study developed a travel speed prediction model based on real-time data collected from the advanced traffic management system (ATMS) in Daegu, South Korea. The model used a clustered K-nearest neighbors (CKNN) algorithm and achieved an average mean absolute percentage error (MAPE) of 6.9% for predicting travel speed in work zones. The predicted travel speed data can be used for optimizing paths and implementing traffic management strategies in work zones.
JOURNAL OF ADVANCED TRANSPORTATION
(2022)
Article
Computer Science, Artificial Intelligence
Renhe Jiang, Zekun Cai, Zhaonan Wang, Chuang Yang, Zipei Fan, Quanjun Chen, Kota Tsubouchi, Xuan Song, Ryosuke Shibasaki
Summary: The density and flow of the crowd or traffic at a citywide level can be predicted using big data and cutting-edge AI technologies. This research topic is significant and has high social impact, as it can be applied to emergency management, traffic regulation, and urban planning. In this study, a new aggregated human mobility dataset is published, and a novel deep learning model called DeepCrowd is proposed for predicting crowd and traffic.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Computer Science, Theory & Methods
Zhaohua Wang, Zhenyu Li, Heng Pan, Guangming Liu, Yunfei Chen, Qinghua Wu, Gareth Tyson, Gang Cheng
Summary: Large cloud service providers have built a growing number of geo-distributed data centers connected by WANs. This article presents a measurement study of high-priority inter-DC traffic from Baidu and proposes a model called IntegNet for traffic prediction. IntegNet leverages both temporal traffic patterns and inferred co-dependencies between DC pairs, leading to significant reductions in QoS losses and overprovisioning.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2023)
Review
Computer Science, Information Systems
Jinshu Su, Baokang Zhao, Yi Dai, Jijun Cao, Ziling Wei, Na Zhao, Congxi Song, Yujing Liu, Yusheng Xia
Summary: This paper summarizes the development trends of network technologies in different fields, including integration, differentiation, and optimization, providing reference for academic researchers and industry professionals to consider future work and build efficient network systems.
FRONTIERS OF INFORMATION TECHNOLOGY & ELECTRONIC ENGINEERING
(2022)
Article
Biochemical Research Methods
Xi Yang, Wei Wang, Jing-Lun Ma, Yan-Long Qiu, Kai Lu, Dong-Sheng Cao, Cheng-Kun Wu
Summary: This paper introduces a deep biological network model BioNet with a graph encoder-decoder architecture for predicting chemical-gene interactions. BioNet utilizes graph convolution to learn latent information from complex interactions among chemicals, genes, diseases and biological pathways. Through parallel training algorithm and multiple GPUs, BioNet achieves outstanding prediction performance in CGI prediction, surpassing current state-of-the-art methods.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Computer Science, Artificial Intelligence
Ge Zheng, Wei Koong Chai, Vasilis Katos
Summary: This paper addresses the problem of multi-step traffic speed prediction and proposes a novel deep learning model, SAGCN-SST, to capture dynamic spatial-temporal processes. Experimental results show that the proposed model consistently achieves the most accurate predictions on two real-world datasets and is robust against emergent traffic situations.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Automation & Control Systems
Teresa Pamula, Renata Zochowska
Summary: In this article, a new method for predicting OD matrix based on traffic data using deep learning is proposed. The method eliminates the need for complex data acquisition and processing, and achieves high accuracy and resistance to missing data. A case study conducted in a medium-sized city in Poland demonstrates the practical implementation potential in real-time traffic assignment systems. The method does not require questionnaire research or detailed spatial development information.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Engineering, Aerospace
Dong Sui, Kechen Liu, Qian Li
Summary: In this study, a spatiotemporal graph convolutional network (STGCN) model is developed to predict air traffic situations. By analyzing time series data, such as sector operational situation and traffic volume, the model accurately predicts the operational situations of multiple sectors, and its effectiveness is proven in the experiments.
Article
Computer Science, Information Systems
Tong Xia, Junjie Lin, Yong Li, Jie Feng, Pan Hui, Funing Sun, Diansheng Guo, Depeng Jin
Summary: The article introduces the 3-Dimensional Graph Convolution Network (3DGCN) framework for predicting citywide crowd flow, achieving superior performance compared to state-of-the-art baselines. By modeling dynamic spatio-temporal graph prediction problems and learning urban structures, the accuracy of predictions is enhanced.
ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA
(2021)
Article
Physics, Multidisciplinary
Weibin Zhang, Huazhu Zha, Shuai Zhang, Lei Ma
Summary: This study proposes a new method for traffic flow prediction by dividing urban road sections into microscopic traffic systems and combining model-driven methods with machine learning. It introduces a traffic flow prediction framework based on the traffic factor state network, which can improve the accuracy of traffic flow predictions.
PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS
(2023)