Spatio-temporal graph mixformer for traffic forecasting

Lablack, Mourad; Shen, Yanming

doi:10.1016/j.eswa.2023.120281

Cited by 19 publications

(5 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most studies in the field of transportation rely on gated linear Unit (GLU) 34 , or gated recursive units (GRU) 35 to capture the dynamic temporal correlation of time series data. Moreover, based on the transformer architecture, STGM 36 introduces a novel attention mechanism to capture both long-term and short-term temporal dependencies. Temporal convolutional networks (TCNs) also have significant advantages in addressing temporal dependencies, especially in time series prediction tasks.…”

Section: Related Workmentioning

confidence: 99%

Graph autoencoder with mirror temporal convolutional networks for traffic anomaly detection

Ren,

Li,

Peng

et al. 2024

Sci Rep

View full text Add to dashboard Cite

Traffic time series anomaly detection has been intensively studied for years because of its potential applications in intelligent transportation. However, classical traffic anomaly detection methods often overlook the evolving dynamic associations between road network nodes, which leads to challenges in capturing the long-term temporal correlations, spatial characteristics, and abnormal node behaviors in datasets with high periodicity and trends, such as morning peak travel periods. In this paper, we propose a mirror temporal graph autoencoder (MTGAE) framework to explore anomalies and capture unseen nodes and the spatiotemporal correlation between nodes in the traffic network. Specifically, we propose the mirror temporal convolutional module to enhance feature extraction capabilities and capture hidden node-to-node features in the traffic network. Morever, we propose the graph convolutional gate recurrent unit cell (GCGRU CELL) module. This module uses Gaussian kernel functions to map data into a high-dimensional space, and enables the identification of anomalous information and potential anomalies within the complex interdependencies of the traffic network, based on prior knowledge and input data. We compared our work with several other advanced deep-learning anomaly detection models. Experimental results on the NYC dataset illustrate that our model works best compared to other models for traffic anomaly detection.

show abstract

Section: Related Workmentioning

confidence: 99%

Graph autoencoder with mirror temporal convolutional networks for traffic anomaly detection

Ren,

Li,

Peng

et al. 2024

Sci Rep

View full text Add to dashboard Cite

show abstract

“…This is due to the CNN reducing the dimensions of the data, keeping important details and avoiding irrelevant ones. Therefore, if CNN layers are used too many times in the problem, important patterns are lost, thus resulting in forecast failure [73]. This characteristic allows the CNN to simplify the problem, making in contrast to the RNN, the training more efficient [74].…”

Section: Temporal Convolutional Networkmentioning

confidence: 99%

Load Forecasting with Machine Learning and Deep Learning Methods

et al. 2023

View full text Add to dashboard Cite

Characterizing the electric energy curve can improve the energy efficiency of existing buildings without any structural change and is the basis for controlling and optimizing building performance. Artificial Intelligence (AI) techniques show much potential due to their accuracy and malleability in the field of pattern recognition, and using these models it is possible to adjust the building services in real time. Thus, the objective of this paper is to determine the AI technique that best forecasts electrical loads. The suggested techniques are random forest (RF), support vector regression (SVR), extreme gradient boosting (XGBoost), multilayer perceptron (MLP), long short-term memory (LSTM), and temporal convolutional network (Conv-1D). The conducted research applies a methodology that considers the bias and variance of the models, enhancing the robustness of the most suitable AI techniques for modeling and forecasting the electricity consumption in buildings. These techniques are evaluated in a single-family dwelling located in the United States. The performance comparison is obtained by analyzing their bias and variance by using a 10-fold cross-validation technique. By means of the evaluation of the models in different sets, i.e., validation and test sets, their capacity to reproduce the results and the ability to properly forecast on future occasions is also evaluated. The results show that the model with less dispersion, both in the validation set and test set, is LSTM. It presents errors of −0.02% of nMBE and 2.76% of nRMSE in the validation set and −0.54% of nMBE and 4.74% of nRMSE in the test set.

show abstract

“…Graph Neural Networks (GNNs) are popularly used in state-of-the-art forecasting models [9,12,24,26,29,40,44,45,50,52,58]. These methods typically represent the traffic sensor network as a graph structure, whose adjacency matrix aims to capture spatial relationships between the sensors.…”

Section: Flow Aggregated Adjacency Matrixmentioning

confidence: 99%

“…Many DL-based forecasting models (and systems such as Foresight) are usually evaluated on benchmark datasets such as METR-LA [23], and seek to forecast 1 hour ahead [8,24,26,29,31,44,57]. While this is useful in short-term forecasting applications, it would be beneficial for transport managers (and members of the public) to be able to extend the forecasting horizons further into the future.…”

Section: Foresight Plus Enhancements 71 Extending the Forecasting Scalementioning

confidence: 99%

“…A wide range of forecasting approaches have been applied to the traffic prediction task, from statistical methods such as ARIMA [28], to Deep Learning (DL) models such as LSTM [18]. In recent years, Graph Neural Network (GNN) approaches have achieved state-of-the-art results, due to their ability to capture spatial dependencies between sensors [17,24,26,29,31,44,52,57]. GNNs typically model the road sensor network as a graph structure, whose weighted adjacency matrix reflects the strength of inter-sensor relationships.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Foresight Plus: Scalable Serverless Real-Time Spatio-Temporal Traffic Forecasting

Oakley,

Conlan,

Demirci

et al. 2023

Preprint

View full text Add to dashboard Cite

Building a real-time, cost-effective, spatio-temporal forecasting system is a challenging problem with many practical applications such as traffic and road network management. Most forecasting research focuses on average prediction quality, with less attention paid to building practical pipelines and achieving timely and accurate forecasts when the network is under heavy load. Additionally, transport authorities need to leverage dynamic data sources (e.g., scheduled roadworks) and vehicle-level flow data, while also supporting ad-hoc inference workloads at low cost. The cloud-based system Foresight, developed in collaboration with Transport for the West Midlands (TfWM), is able to ingest, aggregate and process streamed traffic data, as well as dynamic urban events/flow data to produce regularly scheduled forecasts with high accuracy. In this work, we extend our system with several novel enhancements. First, we present an efficient method for extending the forecasting scale, enabling transport managers to predict traffic patterns further into the future than existing methods. In addition, we augment the existing inference architecture with a new, fully serverless design. This offers a more cost-effective inference solution, which seamlessly handles sporadic inference workloads over multiple forecasting models. We observe that Graph Neural Network (GNN) forecasting models are robust to extensions of the forecasting scale, achieving consistent (and sometimes even improved) performance up to 24 hours ahead. This is in contrast to the 1 hour forecasting horizons popularly considered in the literature. Further, our serverless inference solution is shown to be significantly more cost-effective than provisioned alternatives in appropriate use-cases. We identify the optimal memory configuration of serverless resources to achieve an attractive cost-to-performance ratio.

show abstract

Spatio-temporal graph mixformer for traffic forecasting

Cited by 19 publications

References 17 publications

Graph autoencoder with mirror temporal convolutional networks for traffic anomaly detection

Graph autoencoder with mirror temporal convolutional networks for traffic anomaly detection

Load Forecasting with Machine Learning and Deep Learning Methods

Foresight Plus: Scalable Serverless Real-Time Spatio-Temporal Traffic Forecasting

Contact Info

Product

Resources

About