Real-world Datasets Research Articles

A generative-adversarial-network-based temporal raw trace data augmentation framework for fault detection in semiconductor manufacturing

In modern semiconductor manufacturing, where sophisticated process control mechanisms are standard, processing tools are equipped with sensors that generate vast amounts of raw trace data for process monitoring and fault detection. However, one of the major challenges data scientists face is the scarcity of sufficient raw trace data for defective wafers, creating an imbalance that complicates the training of machine learning models for effective fault detection. To address this issue, this paper proposes novel data augmentation structures and strategies utilizing Cycle Generative Adversarial Networks (CycleGANs) as an artificial intelligence application to synthesize temporal raw trace data for defective wafers. The effectiveness of these methods is demonstrated using a real-world dataset from the thin-film process in semiconductor fabrication. Several machine learning classification models—Gaussian Naive Bayes, Adaptive Boosting, eXtreme Gradient Boosting, and Light Gradient Boosting Machine—are employed to evaluate the performance of the augmented data. The paper identifies the optimal augmentation structure and strategy to enhance classification performance within the CycleGAN-based framework. For the thin-film processing dataset under study, the best classification performance achieves an accuracy rate of up to 99.30%, with a notably low false negative rate of 6.45%.

Enhancing financial risk prediction with symbolic classifiers: addressing class imbalance and the accuracy–interpretability trade–off

Machine learning for financial risk prediction has garnered substantial interest in recent decades. However, the class imbalance problem and the dilemma of accuracy gain by loss interpretability have yet to be widely studied. Symbolic classifiers have emerged as a promising solution for forecasting banking failures and estimating creditworthiness as it addresses class imbalance while maintaining both accuracy and interpretability. This paper aims to evaluate the effectiveness of REMED, a symbolic classifier, in the context of financial risk management, and focuses on its ability to handle class imbalance and provide interpretable decision rules. Through empirical analysis of a real-world imbalanced financial dataset from the Federal Deposit Insurance Corporation, we demonstrate that REMED effectively handles class imbalance, improving performance accuracy metrics while ensuring interpretability through a concise and easily understandable rule system. A comparative analysis is conducted against two well-known rule-generating approaches, J48 and JRip. The findings suggest that, with further development and validation, REMED can be implemented as a competitive approach to improve predictive accuracy on imbalanced financial datasets without compromising model interpretability.

Relieving popularity bias in recommendation via debiasing representation enhancement

The interaction data used for training recommender systems often exhibit a long-tail distribution. Such highly imbalanced data distribution results in an unfair learning process among items. Contrastive learning alleviates the above issue by data augmentation. However, it lacks consideration of the significant disparity in popularity between items and may even introduce false negatives during the data augmentation, misleading user preference prediction. To address this issue, we combine contrastive learning with a weighted model for negative validation. By penalizing identified false negatives during training, we limit their potential harm within the training process. Meanwhile, to tackle the scarcity of supervision signals for unpopular items, we design Popularity Associated Modeling to mine the correlation among items. Then we guide unpopular items to learn hidden features favored by specific users from their associated popular items, which provides effective supplementary information for their representation modeling. Extensive experiments on three real-world datasets demonstrate that our proposed model outperforms state-of-the-art baselines in recommendation performance, with Recall@20 improvements of 4.2%, 2.4% and 3.6% across the datasets, but also shows significant effectiveness in relieving popularity bias.

Time-aware cross-domain point-of-interest recommendation in social networks

Point-of-Interest recommendation within the single domain is quite easy compared to the cross-domain recommendation, as there is an acute dearth of check-in records for the target regions, aggravating the cold start problem. We propose a self-ensembled contextual Thompson sampling for cross-domain Point-of-Interest recommendation to solve this. This approach utilizes user preference transfer and user preference drift in the target domain for enhanced recommendation by deploying enhanced contextual sampling. As the user-context pairs of the target domain are not labeled for the given user, domain adaptation is highly sought. The approach has four major steps: i) Mining Point-of-Interests based on the long-term preferences of the target user and the user with a similar trajectory in the source domain, ii) Computing rewards for the Point-of-Interests in the source domain using multi-layer perceptron, iii) Estimate the rewards for unlabeled Point-of-Interests in the target domain and iv) Form the ensemble of rewards that are used to decide the final arm pulls. The rewards obtained for Point-of-Interests in the source and target domain are combined to form an ensemble of rewards with the help of self ensembling domain adaptation technique. Each Point-of-Interest in the ensemble rewards is termed an arm of action. We use this ensemble of rewards to control the diversity measure and the switching probability of the various arms, potential Point-of-Interests, in the contextual Thompson Sampling. Contextual Thompson sampling is modified to incorporate exploitation-exploration tradeoffs using this reward ensemble. The implicit weight measure of the different arms decides the probability of exploitation or exploration. The final arm pulls results in the final Point-of-Interest recommendation. For experimentation, we have used two real-world datasets, namely, Gowalla and Foursquare, and extracted the data for seven domains. We have obtained an accuracy of approximately 65% for Point-of-Interest recommendations on cold-start users.

Generating Global and Local Explanations for Tree-Ensemble Learning Methods by Answer Set Programming

Abstract We propose a method for generating rule sets as global and local explanations for tree-ensemble learning methods using answer set programming (ASP). To this end, we adopt a decompositional approach where the split structures of the base decision trees are exploited in the construction of rules, which in turn are assessed using pattern mining methods encoded in ASP to extract explanatory rules. For global explanations, candidate rules are chosen from the entire trained tree-ensemble models, whereas for local explanations, candidate rules are selected by only considering rules that are relevant to the particular predicted instance. We show how user-defined constraints and preferences can be represented declaratively in ASP to allow for transparent and flexible rule set generation, and how rules can be used as explanations to help the user better understand the models. Experimental evaluation with real-world datasets and popular tree-ensemble algorithms demonstrates that our approach is applicable to a wide range of classification tasks.

RCCNet: A Spatial-Temporal Neural Network Model for Logistics Delivery Timely Rate Prediction

In logistics service, the delivery timely rate is a key experience indicator, which is highly essential to the competitive advantage of express companies. Prediction on it enables intervention on couriers with low predicted results in advance, thus ensuring employee productivity and customer satisfaction. Currently, few related works focus on couriers’ level delivery timely rate prediction, and there are complex spatial correlations between couriers and road districts in the express scenario, which makes traditional real-time prediction approaches hard to utilize. To deal with this, we propose a deep spatial-temporal neural network, RCCNet to model spatial-temporal correlations. Specifically, we adopt Node2vec, which can encode the road network-based graph directly to capture spatial correlations between road districts. Further, we calculate couriers’ historical time-series similarity to build a graph and employ graph convolutional networks to capture the correlation between couriers. We also leverage historical sequential information with long short-term memory networks. We conduct experiments with real-world express datasets. Compared with other competitive baseline methods widely used in industry, the experiment results demonstrate its superior performance over multiple baselines.

IEDSFAN: information enhancement and dynamic-static fusion attention network for traffic flow forecasting

Accurate forecasting of traffic flow in the future period is very important for planning traffic routes and alleviating traffic congestion. However, traffic flow forecasting still faces serious challenges. Most of the existing traffic flow forecasting methods are static graph convolutional networks based on prior knowledge, ignoring the special spatial–temporal dynamics of spatial–temporal data. Using only adaptive dynamic graphs completely discards the objective and real spatial connectivity information in static graphs. To this end, we propose a novel information enhancement and dynamic-static fusion attention network (IEDSFAN). Firstly, the Multi-Graph Fusion Gating mechanism (MGFG) designed in IEDSFAN effectively fuses dynamic and static graphs to dynamically capture the hidden spatial–temporal correlation. Secondly, we construct a novel Gated Multi-head Self-Attention (GMHSA), which maps the input through the MGFG module to capture the complex spatial–temporal interactions in the features. Finally, we generate adaptive parameters to solve the problem that shared parameters cannot learn multiple traffic patterns, and enhance the expression of sequence information through the peak flag module. We conducted extensive experiments on five real-world traffic datasets, and the experimental results show that the performance of IEDSFAN is significantly better than all baselines.

Abstract B046: Methylation based circulating tumor DNA serial monitoring correlates with immunotherapy response in non-small cell lung cancer

Abstract Purpose: Circulating tumor DNA (ctDNA) can reflect the genetic and epigenetic composition of malignancies and can serve as a non-invasive biomarker for cancer diagnostics and monitoring. This study aimed to evaluate the utility of a methylation based ctDNA assay as a predictive tool in non-small cell lung cancer (NSCLC) anti-PD1 based immunotherapy monitoring. Methods: We evaluated a cohort of patients with NSCLC treated with anti-PD1 based immunotherapy, from which 108 blood specimens were collected at baseline and during treatment follow-up. Tumor Methylation Scores (TMS) were measured with an amplicon-based, multiplexed cfDNA assay that utilizes quantitative counting templates (QCTs) in conjunction with next-generation sequencing to count the number of methylated molecules at more than 500 genomic locations that are hypermethylated in cancer tissue. The association between TMS and real-world progression-free survival (rwPFS) on therapy was conducted using Cox proportional hazards model and plotted using the Kaplan-Meier method. Results: The change in TMS measured 4-10 weeks post-treatment initiation strongly correlated with immunotherapy response, as measured by rwPFS (p<0.0001), compared to a weaker correlation of imaging RECIST v1.1 measurements with rwPFS (p=0.55). Furthermore, TMS tracked with tumor burden on therapy in real-world cases. Conclusions: In this real-world dataset of NSCLC patients treated with anti- PD1 immunotherapy regimens, TMS scores measured within a 4-10 week window after treatment initiation can be predictive of response to therapy. Beyond this window, the TMS score can be associated with rwPFS and tumor dynamics. Early evidence suggests that changes in the methylation profile may be informative for monitoring occurrence of new somatic mutations. The cases presented demonstrate the applicability of using TMS for serial therapeutic response monitoring. Citation Format: Angela Hsiao, Brian Woodward, Patrick Ye, Matthew Varga, Kevin Lu, Ghaith Altaie, Naomi Searle, Robb Veins, Sydne Langpap, Zeqian Li, Gary Palmer, Hatim Husain. Methylation based circulating tumor DNA serial monitoring correlates with immunotherapy response in non-small cell lung cancer [abstract]. In: Proceedings of the AACR Special Conference: Liquid Biopsy: From Discovery to Clinical Implementation; 2024 Nov 13-16; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2024;30(21_Suppl):Abstract nr B046.

Dynamic Spatio-Temporal Hypergraph Convolutional Network for Traffic Flow Forecasting

Graph convolutional networks (GCN) are an important research method for intelligent transportation systems (ITS), but they also face the challenge of how to describe the complex spatio-temporal relationships between traffic objects (nodes) more effectively. Although most predictive models are designed based on graph convolutional structures and have achieved effective results, they have certain limitations in describing the high-order relationships between real data. The emergence of hypergraphs breaks this limitation. A dynamic spatio-temporal hypergraph convolutional network (DSTHGCN) model is proposed in this paper. It models the dynamic characteristics of traffic flow graph nodes and the hyperedge features of hypergraphs simultaneously, achieving collaborative convolution between graph convolution and hypergraph convolution (HGCN). On this basis, a hyperedge outlier removal mechanism (HOR) is introduced during the process of node information propagation to hyper-edges, effectively removing outliers and optimizing the hypergraph structure while reducing complexity. Through in-depth experimental analysis on real-world datasets, this method has better performance compared to other methods.

Adaptive Multi-Scale Bayesian Framework for MFL Inspection of Steel Wire Ropes

Magnetic flux leakage (MFL) technology is widely used in steel wire rope (SWR) inspection for non-destructive testing. However, accurate defect characterization requires advanced signal processing techniques to handle complex noise conditions and varying defect types. This paper presents a novel adaptive multi-scale Bayesian framework for MFL signal analysis in SWR inspection. Our approach integrates discrete wavelet transform with adaptive thresholding and multi-scale feature fusion, enabling simultaneous detection of minute defects and large-area corrosion. To validate our method, we implemented a four-channel MFL detection system and conducted extensive experiments on both simulated and real-world datasets. Compared with state-of-the-art methods, including long short-term memory (LSTM), attention mechanisms, and isolation forests, our approach demonstrated significant improvements in precision, recall, and F1 score across various tolerance levels. The proposed method showed superior detection performance, with an average precision of 91%, recall of 89%, and an F1 score of 0.90 in high-noise conditions, surpassing existing techniques. Notably, our method showed superior performance in high-noise environments, reducing false positive rates while maintaining high detection sensitivity. While computational complexity in real-time processing remains a challenge, this study provides a robust solution for non-destructive testing of SWR, potentially improving inspection efficiency and defect localization accuracy. Future work will focus on optimizing algorithmic efficiency and exploring transfer learning techniques for enhanced adaptability across different non-destructive testing (NDT) domains. This research not only advances signal processing and anomaly detection technology but also contributes to enhancing safety and maintenance efficiency in critical infrastructure.

Abstract 4138794: Risk Factors of Cumulative Limb Amputation in Patients with Thromboangiitis Obliterans (Buerger's Disease): Insights from Nationwide Inpatient Sample

Background: Thromboangiitis obliterans (TAO) is a rare, nonatherosclerotic, inflammatory disease that typically onsets in male patients at young age with heavy smoking. TAO commonly affects small and medium-sized arteries in extremities, and up to 1/3 of patients with TAO will experience limb amputation. However, the role of social determinants of health (SDoH) are not well studied in TAO patients with limb amputation in a real-world dataset. Methods: A cross-sectional study was performed in Nationwide Inpatient Sample (NIS) of United States, from calendar year 2016-2021. The diagnosis of diseases was identified by ICD-10-CM codes. Univariable logistic regression analysis was performed to identify characteristics associated with cumulative limb amputation (CLA). Stepwise multivariate logistic regression model was further employed to analyze association between significant characteristics and CLA. Results: In this study, 33.3% (N=432) patients with TAO (N=1297) had CLA. Variables showed significant association with CLA identified by univariable analysis were further included in stepwise multivariate logistic regression model. In the multivariate model, middle-age patients had higher risk of CLA than young patients (OR=1.41, 95% CI=1.00-1.99, p=0.050). Female showed lower CLA risk than male patients (OR=0.53, 95% CI=0.41-0.69, p<0.001). Black patients showed higher CLA risk than White (OR=1.82, 95% CI=1.24-2.68, p=0.002). Compared to Medicare insurance, self-pay (OR=0.51, 95% CI=0.30-0.87, p=0.014) and Private insurance (OR=0.69, 95% CI=0.49-0.97, p=0.035) patients had lower CLA risk. Higher socioeconomic status associated with lower CLA risk, and patients in the 3 rd percentile had significantly lower CLA risk than those in the 1 st percentile (OR=0.68, 95% CI=0.49-0.95, p=0.028). Patients with tobacco smoking showed higher CLA risk (OR=2.11, 95% CI=2.49-3.00, p<0.001). Conclusions: The CLA rate was up to 33.3% among the TAO patients in NIS 2016-2021. SDoH like age, sex, race, insurance type and socioeconomic status showed significant association with CLA risk in TAO patients. Further studies are warranted to explore mechanism, health equity and multidisciplinary care to decrease CLA for patients with TAO.

ProcessGAN: Generating Privacy-Preserving Time-Aware Process Data with Conditional Generative Adversarial Nets

Process data constructed from event logs provides valuable insights into procedural dynamics over time. The confidential information in process data, together with the data’s intricate nature, makes the datasets not sharable and challenging to collect. Consequently, research is limited using process data and analytics in the process mining domain. In this study, we introduced a synthetic process data generation task to address the limitation of sharable process data. We introduced a generative adversarial network, called ProcessGAN, to generate process data with activity sequences and corresponding timestamps. ProcessGAN consists of a transformer-based network as the generator, and a time-aware self-attention network as the discriminator. It can generate privacy-preserving process data from random noise. ProcessGAN considers the duration of the process and time intervals between activities to generate realistic activity sequences with timestamps. We evaluated ProcessGAN on five real-world datasets, two that are public and three collected in medical domains that are private. To evaluate the synthetic data, in addition to statistical metrics, we trained a supervised model to score the synthetic processes. We also used process mining to discover workflows for synthetic medical processes and had domain experts evaluate the clinical applicability of the synthetic workflows. ProcessGAN outperformed the existing generative models in generating complex processes with valid parallel pathways. The synthetic process data generated by ProcessGAN better represented the long-range dependencies between activities, a feature relevant to complicated medical and other processes. The timestamps generated by the ProcessGAN model showed similar distributions with the authentic timestamps. In addition, we trained a transformer-based network to generate synthetic contexts (e.g., patient demographics) that were associated with the synthetic processes. The synthetic contexts generated by our model outperformed the baseline models, with the distributions similar to the authentic contexts. We conclude that ProcessGAN can generate sharable synthetic process data indistinguishable from authentic data. Our source code is available in https://github.com/raaachli/ProcessGAN .

An adaptive decision support system for outpatient appointment scheduling with heterogeneous service times

Appointment scheduling (AS) plays a crucial role in outpatient clinic management. Traditional methods involve patient grouping using pre-defined rules and scheduling based on these groups. However, pre-defined rules may not adequately capture the heterogeneity in patients’ service times (i.e., consultation duration). Advanced machine learning (ML) methods can address individual-level heterogeneity but pose challenges for practical scheduling. To strike a balance, we propose a data-driven AS decision support system, Cluster-Predict-Schedule (CPS), integrating both supervised and unsupervised ML for efficient patient grouping and scheduling. The novelty of CPS lies in its adaptability to service time heterogeneity through a data-driven approach, determining patient groups based on data rather than pre-defined rules. Additionally, CPS includes a generic and efficient algorithm for generating appointment templates adaptable to any number of patient groups. Our system’s efficacy is demonstrated using a real-world dataset. Evaluated by the weighted sum of patient wait times, physician idle time, and overtime, CPS achieves up to 15.0% cost reduction compared to the FCFA (first-call, first-appointment) scheme and over 4.7% savings against the common New/Return classification with traditional sequencing candidate (TSC) rules. In addition, CPS enhances outpatient operational efficiency without compromising fairness.

Detection and Correction of Likert Scale Multiplicative Response-Style Bias

Individuals may differ in their tendency to fully use the range of a Likert scale. Such individual response styles may have an effect on estimates of correlations. It is shown with simulated data that a relatively simple nonlinear latent model can be used to detect and correct (to some extent) for the bias introduced by differences in multiplicative response styles. A real-world dataset is examined to show that the effect may occur in reality and may lead to biased conclusions.

Towards accurate anomaly detection for cloud system via graph-enhanced contrastive learning

As a critical technology, anomaly detection ensures the smooth operation of cloud systems while maintaining the market competitiveness of cloud service providers. However, the resource data in real-world cloud systems is predominantly unannotated, leading to insufficient supervised signals for anomaly detection. Moreover, complicated topological associations existed between cloud servers (e.g., computation, storage, and communication). While acquiring resource information, correlating the system topology is challenging. To this end, we propose the GCAD for cloud system anomaly detection, which integrates data augmentation, GraphGRU, contrastive learning, and reconstruction. First, GCAD constructs positive and negative sample pairs through the masking and Gaussian noise data augmentation. Then, the GraphGRU processes extended temporal graph data, extracting and fusing spatiotemporal features from resource status and system topology. In addition, GCAD introduces linear attention for encoding spatiotemporal representations to capture their global correlation information. The weight parameters of the encoder are optimized using a contrastive learning mechanism. Finally, GCAD utilizes a reconstruction technique to calculate anomaly scores, facilitating the evaluation of the state of the cloud system at each time point. Experimental results indicate that GCAD outperforms state-of-the-art compared methods on two real-world datasets that contain topology information.

OptiFeat: enhancing feature selection, a hybrid approach combining subject matter expertise and recursive feature elimination method

Optimizing the performance of Java Virtual Machines (JVMs) (Sahin et al. in Proc IEEE Int Congr Big Data BigData Congr 410–417, 2016) is crucial for achieving efficient execution of Java applications. Feature selection plays a pivotal role in identifying the most relevant parameters for fine-tuning JVMs, thereby enhancing their overall efficiency. This paper presents a novel hybrid approach that integrates both subject matter expertise and Recursive Feature Elimination (RFE) (Yin et al. in J Big Data 10(1):15, 2023) model to refine feature selection for JVM fine-tuning using machine learning techniques. Traditional feature selection methods often lack the ability to incorporate domain-specific knowledge, resulting in suboptimal selections (Khaire and Dhanalakshmi in J King Saud Univ Comput Inf Sci 34(4):1060–1073, 2022). In contrast, the hybrid approach leverages the expertise of JVM administrators or developers to guide the feature selection process. By integrating domain knowledge into the feature selection pipeline, ensure the inclusion of crucial JVM parameters that may not be captured by automated techniques alone. Furthermore, employed the RFE model, a powerful recursive feature elimination algorithm, to iteratively identify and eliminate irrelevant features from the initial feature set. This iterative process enhances the efficiency of feature selection by systematically pruning less influential parameters, thereby improving the overall performance of the JVM. To validate the effectiveness of the hybrid approach, conducted experiments using real-world JVM datasets and compare the performance of the method against existing feature selection techniques. The results demonstrate that the approach not only achieves superior performance in terms of JVM fine-tuning but also provides insights into the significance of domain expertise in optimizing JVM performance (Menéndez and Bartlett in http://arxiv.org/abs/2310.16510, 2023). It contributes to the field of JVM optimization by proposing a novel hybrid approach that combines subject matter expertise with machine learning-based feature selection techniques. By leveraging both domain knowledge and automated algorithms, the approach offers a comprehensive solution for enhancing feature selection in JVM fine-tuning, ultimately leading to improved performance and efficiency in Java application execution.

Empowering Youth through Open Data: Map-Based Storytelling of Croatian vs. European Youth

Abstract. In the digital age, open data has become a critical asset for fostering transparency, accountability, and civic engagement. However, the widespread availability of data has outpaced citizens' ability to effectively utilize it, underscoring the need for data literacy skills, especially among youth. This paper addresses the gap in data literacy by proposing an innovative teaching approach that leverages map-based storytelling to engage young people and enhance their data literacy. By integrating real-world open datasets into interactive visual narratives, this approach aims to captivate young audiences, develop critical thinking skills, and foster a deeper understanding of socio-economic issues. Through a case study of a Datathon project titled “Croatian Youth vs. EU: Who has it better?”, this research demonstrates the efficacy of combining data visualization and storytelling to promote data literacy and civic engagement. The findings highlight the potential of open data initiatives and open-source technologies in empowering youth and enhancing their participation in open government initiatives and policy-making processes.

Toward Cross-Lingual Social Event Detection with Hybrid Knowledge Distillation

Recently published graph neural networks (GNNs) show promising performance at social event detection tasks. However, most studies are oriented toward monolingual data in languages with abundant training samples. This has left the common lesser-spoken languages relatively unexplored. Thus, in this work, we present a GNN-based framework that integrates cross-lingual word embeddings into the process of graph knowledge distillation for detecting events in low-resource language data streams. To achieve this, a novel cross-lingual knowledge distillation framework, called CLKD, exploits prior knowledge learned from similar threads in English to make up for the paucity of annotated data. Specifically, to extract sufficient useful knowledge, we propose a hybrid distillation method that consists of both feature-wise and relation-wise information. To transfer both kinds of knowledge in an effective way, we add a cross-lingual module in the feature-wise distillation to eliminate the language gap and selectively choose beneficial relations in the relation-wise distillation to avoid distraction caused by teachers’ misjudgments. Our proposed CLKD framework also adopts different configurations to suit both offline and online situations. Experiments on real-world datasets show that the framework is highly effective at detection in languages where training samples are scarce.

Reliability Analysis for Unknown Age Class of Lifetime Distribution with Real Applications in Medical Science

Analyzing the reliability of the aging class of life distribution provides important information about how long a product lasts and sustainability measures that are essential for determining the environmental impact and formulating resource-saving plans. The study emphasizes the goodness-of-fit technique of the nonparametric test for the NBRUmgf class because age data are crucial for applications. Evaluations were conducted using the test’s asymptotic properties and Pitman efficiency methodology for some selected asymmetric probability models, and the outcomes were compared with those of alternative methods. We assessed the test’s power against widely used reliability distributions for some well-known alternative asymmetric distributions, including the Weibull, Gamma, and linear failure rate (LFR) distributions, and provided percentiles for both censored and uncensored data. This study shows the efficacy of the test in various sectors using real-world datasets and comprehensive tables of test statistics.

Spatio-temporal attention based collaborative local–global learning for traffic flow prediction

Traffic flow prediction is crucial for intelligent transportation systems (ITS), providing valuable insights for traffic control, route planning, and operation management. Existing work often separately models the spatial and temporal dependencies and primarily relies on predefined graphs to represent spatio-temporal dependencies, neglecting the traffic dynamics caused by unexpected events and the global relationships among road segments. Unlike previous models that primarily focus on local feature extraction, we propose a novel collaborative local–global learning model (LOGO) that employs spatio-temporal attention (STA) and graph convolutional networks (GCN). Specifically, LOGO simultaneously extracts hidden traffic features from both local and global perspectives. In local feature extraction, a novel STA is devised to directly attend to spatio-temporal coupling interdependencies instead of separately modeling temporal and spatial dependencies, and to capture in-depth spatio-temporal traffic context with an adaptive graph focusing on the dynamics in traffic flow. In global feature extraction, a global correlation matrix is constructed and GCNs are utilized to propagate messages on the obtained matrix to achieve interactions between both adjacent and similar road segments. Finally, the obtained local and global features are concatenated and fed into a gated aggregation to forecast future traffic flow. Extensive experiments on four real-world traffic datasets sourced from the Caltrans Performance Measurement System (PEMS03, PEMS04, PEMS07, and PEMS08) demonstrate the effectiveness of our proposed model. LOGO achieves the best performance over 18 state-of-the-art baselines and the best prediction performance with the highest improvement of 6.06% on the PEMS07 dataset. Additionally, two real-world case studies further substantiate the robustness and interpretability of LOGO.