Traffic prediction based on real-world traffic data is a crucial task in Intelligent Transportation Systems (ITS). However, the issue of missing observations due to real-world disturbances undermines the robustness and accuracy of traffic prediction. This problem necessitates the development of a prediction model that integrates the imputation mechanism to be compatible with missing observations. This paper introduces ATTST, a self-imputation-assisted prediction model specifically designed to address the challenge of missing observations in the traffic prediction task. Unlike traditional approaches utilizing an additional supervised imputation model before prediction, the imputation unit in our model does not need the extra label for the missing observations. Our model employs a self-imputation unit to impute the missing observations by partially masking the observed data as the ground true labels. Thus, the self-imputation unit along with an encoder–decoder architecture and a graph evolving unit together directly predict future traffic data with multi-level missing observations. The effectiveness of ATTST is validated using several real-world traffic datasets, including speed and flow data, across various multi-step prediction scenarios with diverse missing observations. These validations demonstrate the model’s robustness and practical applicability in real-world traffic prediction tasks. The results show that ATTST can reliably predict traffic conditions even with incomplete data, making it a valuable tool for traffic management and planning. • The problem of imputation and prediction for traffic speed and flow data with missing observations is formulated within an end-to-end framework. • The proposed AttSt model addresses this problem by incorporating a self-imputation mechanism to effectively handle missing observations. • The model employs a graph network to capture and process spatiotemporal correlations within the traffic data. • Comprehensive evaluations on four real-world traffic speed and flow datasets validate the effectiveness of the proposed approach.

With the rapid development of Intelligent Transportation Systems (ITSs), traffic prediction, a crucial component of ITSs, has garnered growing scholarly attention. The appli-cation of deep learning into traffic prediction has emerged as a prominent research direction, especially amid the rapid advancement of pretrained large language models (LLMs), which offer substantial benefits in time-series analysis through cross-modal knowledge transfer. In response to this advancement, this study introduces an innovative model for traffic flow prediction, designated as WGLLM. To capture spatiotemporal characteristics inherent in traffic flow data, this model incorporates a sequence embedding layer constructed on the stationary wavelet transform (SWT) and long short-term memory (LSTM), in conjunction with a spatial embedding layer founded on graph convolutional networks (GCNs). Additionally, a fully connected layer is utilized to integrate embeddings into the LLMs for comprehensive global dependency analysis. To verify the effectiveness of the proposed approach, experiments were carried out on two real traffic flow datasets. The experimental results demonstrate that WGLLM achieves superior predictive performance compared to multiple mainstream baseline models, accompanied by a significant enhancement in prediction accuracy.

Traffic flow congestion is a pervasive global phenomenon. Nonetheless, the systematic analysis and identification of traffic flow patterns remain a challenge as the volume of traffic data increases. Consequently, robust data extraction methods are required to uncover underlying data patterns. This paper proposes a 3-Satisfiability logic mining approach using a Discrete Hopfield Neural Network, develops the 3-Satisfiability Reverse Analysis method by integrating the Discrete Hopfield Neural Network with a Genetic Algorithm, and implements this method on traffic flow datasets, comparing its accuracy with existing approaches. The 3-Satisfiability Reverse Analysis method employs 3-Satisfiability for logical representation and integrates a Discrete Hopfield Neural Network with a Genetic Algorithm as its learning system. A simulation was conducted using the Urban Traffic dataset for São Paulo, Brazil. The robustness of the method in extracting relationships within traffic flow data was evaluated using selected performance metrics. The results indicated that the proposed 3-Satisfiability Reverse Analysis method, which integrates the Discrete Hopfield Neural Network and Genetic Algorithm, achieved promising performance with an accuracy rate of 80%, outperforming existing methods

High-quality datasets are increasingly recognized as foundational inputs to economic development, industrial upgrading, and public governance. A rigorous evaluation system for data asset quality is therefore needed to improve data governance and to enable value realization in circulation. Focusing on three representative circulation scenarios—data interaction, data exchange, and data trading—this study develops an indicator system from technical, business, and benefit-oriented dimensions. The Decision-Making Trial and Evaluation Laboratory (DEMATEL) method is used to identify causal relationships among indicators and key drivers. To integrate multi-expert judgments under uncertainty, hesitant linguistic variables and evidence theory are adopted, and the Best–Worst Method (BWM) is applied to derive more consistent indicator weights. The resulting weights are combined with the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) to obtain a comprehensive ranking of data asset quality with scenario-adjustable emphasis. A traffic-flow dataset from a data technology enterprise is used to demonstrate applicability and effectiveness. The proposed framework advances scenario-adaptive data quality evaluation and supports enterprise data governance, data transaction pricing, and the implementation of high-quality dataset policies.

Accurate traffic flow prediction is a core capability of intelligent transportation systems, supporting trip planning, network dispatch, and management decisions. Most existing approaches overlook interactive learning of spatiotemporal dependencies and struggle to accommodate pattern diversity arising from spatial heterogeneity and multi-scale temporal variation. To address these limitations, this paper proposes Spatiotemporal-Decoupled Interactive Learning (STDIL), a framework comprising a spatiotemporal decoupling module and an interactive learning module. The former reconstructs sequences along spatial and temporal dimensions to yield more discriminative contextual representations. The latter dynamically reconstructs the graph structure in a data-driven manner to capture spatiotemporal correlations from global and local perspectives, thereby leveraging both neighborhood information and long-range dependencies. Experiments on four real-world urban traffic flow datasets show that STDIL attains significantly higher accuracy than existing methods across all prediction horizons. These results demonstrate STDIL's effectiveness in handling spatiotemporal heterogeneity and dynamic dependencies, as well as its adaptability to diverse traffic scenarios.

The development of the Internet of Things (IoT) has enhanced smart city services for traffic monitoring, leading to numerous schemes for accurate flow prediction based on traffic sensors. However, existing approaches primarily capture spatio-temporal (ST) dependencies from traffic graphs and train their models using randomly ordered data. This overlooks the fact that the modeling difficulty of each sensor/node in the ST traffic graph can vary significantly due to its spatial dependencies and temporal trends, resulting in unreliable and unstable predictions in IoT scenarios. In this context, we argue that a well-designed curriculum with an easy-to-difficult order can improve the training of ST models. Therefore, this paper introduces an ST difficulty measurer to score the node-level difficulty of traffic graph from both spatial and temporal aspects, and then implements a curriculum in the ST model training process. More specifically, based on the tiered ST difficulty score, the ST model training begins with a subgraph consisting of “easy” nodes characterized by relatively consistent spatial relationships and regular temporal patterns. Gradually, more difficult nodes are incorporated into the subgraph and participate in subsequent training stages. Comprehensive experiments and analysis on two real-world traffic flow datasets confirm the effectiveness of our proposed approach.

Traffic systems are quintessential complex systems, characterized by nonlinear interactions, multiscale dynamics, and emergent spatiotemporal patterns over complex networks. These properties make traffic prediction highly challenging, as it requires jointly modeling stable global topology and time-varying local dependencies. Existing graph neural networks often rely on predefined or static learnable graphs, overlooking hidden dynamic structures, while most RNN- or CNN-based approaches struggle with long-range temporal dependencies. This paper proposes a Spatiotemporal Causal–Trend Network (SCTN) tailored to complex transportation networks. First, we introduce a dual-path adaptive graph learning scheme: a static graph that captures global, topology-aligned dependencies of the complex network, and a dynamic graph that adapts to localized, time-varying interactions. Second, we design a Gated Temporal Attention Module (GTAM) with a causal–trend attention mechanism that integrates 1D and causal convolutions to reinforce temporal causality and local trend awareness while maintaining long-range attention. Extensive experiments on two real-world PeMS traffic flow datasets demonstrate that SCTN consistently achieves superior accuracy compared to strong baselines, reducing by 3.5–4.5% over the best-performing existing methods, highlighting its effectiveness for modeling the intrinsic complexity of urban traffic systems.

This paper compares the effectiveness of support vector machine (SVM), convolutional neural network-long short-term memory (CNN-LSTM), and support vector machine-long short-term memory (SVM-LSTM) models for traffic flow prediction in intelligent transportation systems (ITS). This research aims to explore the application scenarios of different machine learning and neural network models in traffic flow forecasting, focusing on verifying the effectiveness of the CNN-LSTM and the SVM-LSTM models designed in this paper in integrating spatial feature extraction with time series modeling. Experimental validation is conducted on real-world long- and short-term traffic flow datasets, and the performance of each model is systematically evaluated in terms of the number of prediction errors, computational efficiency, and robustness. Through a comprehensive analysis of metrics such as the coefficient of determination (R2) and root mean square error (RMSE), thisresearch provides a basis for the appropriate selection of prediction models in ITS and offers theoretical support for future research in multimodal traffic data fusion modeling. Purpose. Through systematic comparative studies, a more efficient and reliable model is screened out for the traffic flow prediction subsystem in ITS, and the effectiveness of the hybrid model in integrating multi-dimensional features is explored, thus providing an empirical basis for further optimization of model accuracy in the future. Materials and methods. This research used a long-term traffic flow dataset from France and a short-term traffic flow dataset from Italy. Prediction experiments were conducted in the MATLAB environment using support vector machines (SVMs), CNN-LSTM models, and an SVM-LSTM model with a loss function. The method for determining model effectiveness is based on linear regression theory, focusing on calculating the number of error data and evaluating the data fit using metrics. The method for determining model effectiveness is based on linear regression theory, focusing on calculating the number of error data and evaluating the data fit using metrics. Results. Experimental results based on real-world traffic flow datasets show that the SVM-LSTM model exhibits the best overall performance in long-term traffic flow prediction. The CNN-LSTM model demonstrates excellent time series modeling capabilities in short-term traffic flow prediction. In terms of computational efficiency, the SVM-LSTM model improves prediction accuracy by 10.2% compared to the CNN-LSTM model. Therefore, the fusion of SVM and LSTM combines the advantages of spatial feature extraction and time series modeling, and its deployment in ITS can improve traffic flow prediction efficiency.

This paper proposes an Prediction of Multivariate Spatial-Temporal Series Data Based on Adaptive Spatial-Temporal Information(ASTCN), which learns complex spatio-temporal information from multivariate spatio-temporal series through trainable temporal embeddings and graph adjacency matrices. A gated fusion mechanism is employed to control the proportion of different temporal embeddings to improve prediction accuracy. Visualization of the temporal embeddings reveals that weekly temporal embeddings have the greatest impact on prediction accuracy, followed by daily temporal embeddings, while monthly temporal embeddings have the least impact. Additionally, a novel method for constructing graph adjacency matrices is introduced. Ablation experiments demonstrate that the two types of graph adjacency matrices proposed in this method have varying degrees of influence on improving the prediction accuracy of the dataset. Consequently, this paper integrates the two graph adjacency matrices, enabling ASTCN to achieve superior prediction accuracy on traffic speed, traffic flow, and air quality datasets compared to when either matrix is used alone. In comparative experiments, ASTCN ultimately achieves excellent prediction performance with relatively low training costs.

Long-sequence traffic flow forecasting plays a crucial role in intelligent transportation systems. However, existing Transformer-based approaches face a quadratic complexity bottleneck in computation and are prone to over-smoothing in deep architectures. This results in overly averaged predictions that fail to capture the peaks and troughs of traffic flow. To address these issues, we propose a State-Space Generative Adversarial Network (SSGAN) with a state-space generator and a multi-scale convolutional discriminator. Specifically, a bidirectional Mamba-2 model was designed as the generator to leverage the linear complexity and efficient forecasting capability of state-space models for long-sequence modeling. Meanwhile, the discriminator incorporates a multi-scale convolutional structure to extract traffic features from the frequency domain, thereby capturing flow patterns across different scales, alleviating the over-smoothing issue and enhancing discriminative ability. Through adversarial training, the model is able to better approximate the true distribution of traffic flow. Experiments conducted on four real-world public traffic flow datasets demonstrate that the proposed method outperformed the baselines in both forecasting accuracy and computational efficiency.

Traffic flow forecasting task plays an essential role in intelligent transportation systems. Accurately capturing the intricate spatio-temporal dependencies in traffic network signals is the core of precise prediction. Recently, a paradigm that models spatio-temporal dependencies through graph neural networks and time series models has become one of the most promising methods to solve this problem. However, existing methods still have limitations due to ineffectively modeling dynamic spatial dependencies and high time and space complexity. To address these issues, we propose a simplifying and powerful general spatio-temporal traffic flow forecasting model called LightST. Specifically, LightST first embeds temporal covariates and spatial position information to enhance the spatio-temporal modeling capabilities. Then, stacked temporal linear layers are introduced to capture temporal dependencies efficiently. Finally,we propose a concise adaptive spatio-temporal embedding graph convolution method to extract implicit spatial dependencies over time via dynamic graph convolution with adaptive spatio-temporal embedding graph generation. Extensive experiment results on four public traffic flow datasets demonstrate the superiority of our LightST concerning computational efficiency and prediction performance.

Traffic flow prediction is a prominent research area in intelligent transportation systems, significantly contributing to urban traffic management and control. Existing methods or models for traffic flow prediction predominantly rely on a fixed-graph structure to capture spatial correlations within a road network. However, the fixed-graph structure can restrict the representation of spatial information due to varying conditions such as time and road changes. Drawing inspiration from the attention mechanism, a new prediction model based on the mixed-graph neural network is proposed to dynamically capture the spatial traffic flow correlations. This model uses graph convolution and attention networks to adapt to complex and changeable traffic and other conditions by learning the static and dynamic spatial traffic flow characteristics, respectively. Then, their outputs are fused by the gating mechanism to learn the spatial traffic flow correlations. The Transformer encoder layer is subsequently employed to model the learned spatial characteristics and capture the temporal traffic flow correlations. Evaluated on five real traffic flow datasets, the proposed model outperforms the state-of-the-art models in prediction accuracy. Furthermore, ablation experiments demonstrate the strong performance of the proposed model in long-term traffic flow prediction.

Traffic flow prediction is the key to accurate urban traffic control and the basis for developing intelligent transportation systems. Recent studies have made substantial progress in traffic prediction by modelling complex spatiotemporal graph topology and considering sensors as road network nodes. However, the current spatiotemporal graph neural network model is limited by its structure. It can only utilize short-range traffic flow data and cannot effectively extract the long-term trend of complex traffic flow and periodic features in traffic patterns. To address the above problems, we propose a Transformer-based long-term traffic flow prediction framework, “Transformer-based spatiotemporal graph attention network”. First, the model utilizes the Transformer coding layer to learn compressed and context-rich subsequence temporal representations from long-term sequences. Then, the model designs a multi-scale gated temporal convolution module to identify and extract long-term trend features of traffic flow from the subsequence time representations. Next, the model constructs a multi-granularity random graph attention module to capture the periodic features of traffic flow from the subsequence time representations and extracts the short-term trend features present in the long-time series using the STGNN model. Finally, the model fuses the extracted long-term trends, periodic features and short-term trends to obtain the final prediction results. Experimental results on two real-world traffic flow datasets show that the model outperforms the baseline model and makes accurate long-term predictions.

The emergence of the Internet of Things (IoT) has introduced a variety of devices into smart homes, making smart home networks increasingly complex and insecure. However, many IoT device manufacturers prioritize functionality, time-to-market, and performance over security, leaving IoT devices and networks vulnerable. Automatic device classification techniques are crucial for applying various network management approaches to ensure both performance and security. Despite the considerable research effort devoted to device classification, very few datasets are publicly available for in-depth investigation. This paper identifies the currently available public datasets for smart home device classification and highlights their limitations. These limitations encouraged us to develop a new, large-scale network traffic flow dataset for AI-Based smart home device classification dataset comprising more than 200 million data points stemming from 105 different IoT and non-IoT devices. This dataset is now publicly available to the research community, and in this paper we present and describe its properties. Furthermore, we evaluated the effectiveness of different Machine Learning algorithms in classifying these devices. Our results indicate that the Random Forest algorithm achieves the highest accuracy at 0.906 with recall, precision, and F1 scores of 0.877, 0.901, and 0.887, respectively. Finally, we investigated the importance of the features and found that only 12 features are largely responsible for the observed levels of accuracy. • Current datasets for smart home device classification have limited use in practice. • We developed and characterized the largest dataset to date with > 200 M datapoints. • A key property of the dataset is its diversity. • The dataset has 88 features from 105 devices spanning 23 different types. • Random Forest performs best on this dataset with metrics around 90%.

Abstract As an effective means to solve the challenges of online measurement of difficult-to-measure variables in complicated traffic processes, deep learning–based traffic flow prediction has emerged as one of the primary research objectives in the field of soft measurement with its strong data feature extraction ability. However, most of the existing traffic flow prediction methods follow a stacked structure of spatial-temporal blocks to capture traffic flow features, which encounter limitations in solving the over-smoothing problem caused by the deep-stacked graph convolution network (GCN), as well as the interaction effects endured by the spatial-temporal attention learning based on cascaded structure. In this paper, we adopt a sequential structure to connect the spatial module, the temporal module, and the attention module to achieve the spatial-temporal correlations learning. Along this line, we propose a deep graph convolution spatial-temporal attention learning network, which considers both large range spatial dependence and global joint spatial-temporal correlation, to predict traffic flows. In particular, a deep stacked GCN module is adopted in our model to capture multiscale spatial features of the traffic flow data, which are constructed by leveraging a combination of residual networks, jump connections and multilayer perceptrons. Afterward, the temporal features are learned using the gated temporal convolution. Finally, a spatial-temporal attention mechanism is introduced to simultaneously capture dynamic global correlations in both spatial and temporal dimensions. Our experimental results, based on four real-world traffic flow datasets called PeMSD3, PeMSD4, PeMSD7, and PeMSD8, reveal that the prediction accuracy of the proposed model outperforms other baseline models. Moreover, it can efficiently alleviate over-smoothing effects while maintaining manageable computational overhead.

Balancing the need to satisfy both long-term and short-term requirements and comprehensively considering spatial and temporal dependencies are key challenges in metro passenger prediction. A trend spatio-temporal adaptive graph convolution network (TSTA-GCN) model for metro passenger flow prediction is presented in this paper. A trend convolutional self-attention model is designed to learn long-term and short-term trends. Adaptive graph is utilized to capture the complex relationships between stations and an adaptive graph convolutional recurrent unit module is proposed to capture local spatial and dynamic spatio-temporal correlations. In order to simulate the spatio-temporal heterogeneity implied in traffic flow, a spatio-temporal interaction module is used to fuse the heterogeneity in space and time. Extensive experiments are carried out on two metro traffic flow datasets and the experimental results show that the TSTA-GCN model outperforms the state-of-the-art baseline methods and is able to effectively predict long-term and short-term metro passenger flow.

This paper proposes a traffic flow prediction model based on the fusion of time-domain convolutional network (TCN) and long-short-term memory network (LSTM), and verifies its effectiveness through multi-dimensional experiments. To address the complexity of urban traffic flow prediction, the study constructs a TCN-LSTM hybrid model to fuse temporal feature extraction and long and short-term dependency capturing capabilities, and performs prediction validation on three types of traffic flow datasets, namely, cars, bicycles and trucks, respectively. The experimental results show that: in car traffic prediction, the training loss value of the model decreases significantly from the initial 0.8 to less than 0.1, and the R of the training set and the test set reaches 0.73 and 0.75, respectively, which reflects good convergence and generalisation ability; bicycle traffic prediction shows that the R of the training set reaches as high as 0.84 but the R of the test set decreases to 0.31, which shows that there is a certain degree of overfitting phenomenon; truck Truck traffic prediction achieves a balanced performance of R 0.73 for the training set and R 0.50 for the test set, which verifies the model's ability to capture heavy vehicle traffic patterns robustly. By comparing the performance differences between TCN, LSTM and their hybrid models through ablation experiments, it is found that TCN-LSTM is superior in key indicators: the mean absolute error (MAE) is 0.11 lower than that of the pure TCN model, the mean squared error (MSE) is in between that of TCN and LSTM, and the relative prediction deviation (RPD) reaches the highest value of the three at 2.15, which is a good proof that the hybrid model combines both multi-scale capturing ability of time-series features and the advantage of accurate modelling of nonlinear relationships.

Real-time and accurate short-term traffic flow forecasting can provide important decision support for traffic guidance and management. To effectively address the spatial–temporal feature mining problem in short-term traffic flow prediction for complex road networks, a new method that combined the Pearson correlation coefficient (PCC) and improved genetic algorithm to optimize the long short-term memory model (IGA-LSTM) was constructed. It filters the traffic flow data of roads which are related to the spatial characteristics of the target road in the road network through the PCC model, and then the traffic flow data set was reconstructed. Secondly, the new traffic flow data set is treated as the input of the IGA-LSTM, so the PCC-IGA-LSTM forecasting model was proposed that can evaluate the influence of relevant roads on the target road. Finally, the performance of the model was evaluated with Seattle traffic flow data, and experiments of 5-min short-term traffic flow forecasting on both weekdays’ and weekends’ data sets verified the performance of the model respectively. Compared with the other forecasting models such as the IGA-LSTM, the PSO-BP the GA-BP model, and the proposed PCC-IGA-LSTM model, the experimental results show that the proposed model can better integrate the spatial–temporal correlation in the traffic flow data, and the accuracy is improved.

ABSTRACT With the constant dynamics of temporal dependence and spatial correlation, the interaction between them has become intricate. Existing work attempts to model precise temporal dependency and spatial correlation to make their interactions more accurate but ignores the importance of understanding how the two interact with each other. Thus, this article mines deeper into their interaction mechanism and proposes a new traffic prediction model called traffic flow prediction model based on spatial–temporal identity (TPST). It provides a new way named the spatial–temporal identity mechanism to model spatial–temporal interactions, which convert complex temporal dependence and spatial correlation into their identity information. Meanwhile, in order to improve spatial–temporal interaction resolution of the model, the method utilizes the down‐sampling cross‐convolution technique to contain more spatial–temporal history information and parses spatial–temporal interactions at different granularity. Experiments conducted with four real traffic flow datasets show that TPST consistently outperforms the other seven benchmark models, providing higher prediction accuracy with lower computational cost.

Traffic flow data has been used in various disciplines, including geography, transportation, urban planning, and public health. However, existing datasets often have limitations such as low spatiotemporal resolution and inconsistent quality due to data collection methods and the need for an adequate data cleaning process. This paper introduces a long-term traffic flow dataset at an intra-city scale with high spatio-temporal granularity. The dataset covers the Glasgow City Council area for four consecutive years spanning the COVID-19 pandemic, from October 2019 to September 2023, providing comprehensive temporal and spatial coverage. Such detailed information facilitates diverse applications, including traffic dynamic analysis, traffic management, infrastructure planning, and urban environment improvement. Also, it provides a valuable dataset to understand traffic flow change during a once-in-a-lifetime pandemic event.