Pith. sign in

REVIEW 4 major objections 6 minor 54 references

STARN-GAT: A Multi-Modal Spatio-Temporal Graph Attention Network for Accident Severity Prediction

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read STARN-GAT, a graph attention network that fuses road topology, timing, and weather, predicts accident severity with a Macro F1 of 0.85 on US crash data and 0.84 on a Bangladesh dataset, a 2.4-point gain over the strongest recent baseline.

desk verdict A plausible architecture with a thorough write-up, but the FARS evaluation contradicts the dataset definition and test-split arithmetic, so the SOTA claims are unsupported. read the letter →

arxiv 2507.20451 v1 pith:7PWRRFV6 submitted 2025-07-28 cs.AI

Pritom Ray Nobin , Imran Ahammad Rifat This is my paper · ORCID
classification cs.AI
keywords accidentseveritypredictiongraphattentionnetworkspatio-temporalmodelingmulti-modalfusionroadsafetyFARSdatasetARI-BUETfocalloss
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

STARN-GAT is proposed as a way to predict how severe a traffic accident will be by treating roads as a graph and letting attention decide which spatial, temporal, and environmental clues matter for each crash. The paper's central claim is that this architecture reaches a Macro F1 of 0.85 on the FARS data and 0.84 on the ARI-BUET data, a 2.4-point improvement over the strongest recent baseline (ST-GTrans) on both datasets. If the claim holds, emergency response and road-safety planning could use a model that catches 81% of severe accidents in the US data and 78% in the Bangladesh data, precisely the cases that are rare but most consequential. The paper further argues, on the basis of ablations, that the spatial graph attention is the most important component, followed by temporal encoding and external context.

What carries the argument

The load-bearing mechanism is the attention-based fusion of four modules: an adaptive graph built with DBSCAN clustering and a weighted similarity function over topological connectivity, spatial proximity, and functional similarity; a four-head graph attention encoder with edge-feature-aware coefficients, residual connections, and concatenated heads; separate temporal and external encoders that map 11 time features and 8 contextual features into 64-dimensional vectors; and a self-attention fusion layer that treats the three 64-dimensional modality embeddings as a 3 by 64 query matrix and applies scaled dot-product attention, then flattens to 192 dimensions for the classifier. Focal loss with gamma equal to 2 is what makes the rare severe class learnable, and cosine annealing with warm restarts plus gradient clipping is the training schedule that carries the reported numbers.

What would settle it

Inspect the actual FARS 2018–2020 file and count records by severity class: if 'no injury' and 'minor' labels are absent, or their counts fall far short of the 40,228 and 28,579 reported in Table IV, then the four-class training described in the paper could not have happened and the FARS metrics are unreproducible.

Watch

Extended reading notes

Core claim

The discovery claimed is that accident severity classification improves when the road network, the crash's timing, and the surrounding context are fused through attention rather than concatenated. Concretely, STARN-GAT constructs a weighted directed graph whose edges encode topological, proximity, and functional similarity between road segments; encodes node features with four-head graph attention; encodes cyclical time features through a two-layer network; and combines the three modality embeddings with a scaled dot-product self-attention layer before a three-layer classification head. On this design the paper reports Macro F1 of 0.85 on FARS and 0.84 on ARI-BUET, severe-class recall of 0.81 and 0.78, and ROC-AUC of 0.91 and 0.89, with McNemar's test p=0.004 separating it from the best baseline. The authors also report that removing the graph attention layer is the single most damaging ablation, cutting Macro F1 by 0.09, and that focal loss raises severe recall from 0.72 to 0.81 on FARS.

Load-bearing premise

The whole evaluation rests on the assumption that the FARS records used in the experiments carry authentic labels for all four severity classes—no injury, minor, moderate, and severe—in roughly the proportions shown in Table IV; if the primary file contains only fatal crashes, as Section V.A describes, the headline FARS results and the 2.4-point improvement do not stand.

Editorial extensions

If this is right

  • If the reported results are correct, a practical accident-severity model can be built from publicly available crash records plus road-network, weather, and time features, reaching Macro F1 of 0.85 and 0.84 without proprietary data.
  • The ablation order (graph attention, then temporal encoding, then external context, then fusion) gives deployment designers a priority list: invest first in a faithful road-network graph, then in time features, then in weather context.
  • The claimed 125 ms inference time and linear scaling law (processing time equals 0.028 times network size plus 17.2 milliseconds) imply the model could be embedded in real-time traffic-management pipelines at city scale.
  • The severe-recall gain from focal loss (0.72 to 0.81 on FARS) means the same loss choice could be imported into other imbalanced safety classifiers, not just this architecture.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial extension: the FARS four-class premise is directly checkable, since the paper itself describes FARS as a record of fatal crashes; a reader should verify that the downloaded FARS file actually contains 'no injury' and 'minor' labels before trusting the 0.85 FARS number.
  • Editorial extension: if re-run on a dataset that genuinely mixes all severities, the relative gain over ST-GTrans could shrink or grow, and the 2.4-point margin is only as meaningful as the label distribution it was computed on.
  • Editorial extension: the attention-fusion module is task-agnostic and could be transferred to other spatio-temporal risk-classification problems, such as emergency triage or infrastructure failure risk, wherever graph, time, and context modalities coexist.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper introduces STARN-GAT, a spatio-temporal graph attention network for multi-class accident severity prediction. The model combines adaptive road-network graph construction, multi-modal feature encoding (spatial, temporal, external context), an attention-based fusion mechanism, and a focal-loss-trained classification head. The central empirical claim is that STARN-GAT outperforms four recent graph-based baselines on the FARS and ARI-BUET datasets, with Macro F1 improvements of roughly 2.4% over the best baseline ST-GTrans (0.85 vs. 0.83 on FARS; 0.84 vs. 0.82 on ARI), severe recall of 0.81 (FARS) and 0.78 (ARI), and ROC-AUC of 0.91 and 0.89. The paper also reports ablation studies, temporal performance dynamics, class-specific analysis, and computational scalability.

Significance. If the reported results were valid, STARN-GAT would represent a solid empirical contribution to graph-based traffic accident severity prediction, combining several architectural ideas (adaptive multi-criteria graph construction, modality-aware attention fusion, focal loss) and evaluating them on two real-world datasets. The paper is reasonably detailed in its architecture description and includes ablation and computation analyses. However, the significance is substantially undercut by a load-bearing data-integrity problem: the FARS dataset is described as a fatal-crash-only database, yet Table IV reports a four-class distribution with 77.3% non-fatal cases. In addition, the support counts in Table IV sum to the full dataset rather than to the stated 15% test split, and no variance/error-bar information is provided despite claims of 5-fold cross-validation. These issues place the main empirical claims in doubt.

major comments (4)
  1. [Section V.A, Table IV] The description of FARS as 'a detailed record of fatal crashes in North America' is incompatible with the four-class severity labels in Table IV, where 45.2% of cases are 'No Injury' and 32.1% are 'Minor'. A fatal-crash database cannot contain crash-level outcomes of no injury or minor injury. If the unit of analysis is instead involved persons, then the problem formulation (accident severity), the graph nodes (road segments), and the prediction target do not match the data unit. This contradiction puts the validity of the entire FARS experiment in question.
  2. [Table IV, Section V.D] The Support column in Table IV sums to 89,010 (40,228 + 28,579 + 15,842 + 4,361), which is approximately the full 89,720-record dataset, not the 15% test split (about 13,458 records) described in Section V.D.1. Consequently, the reported Macro F1 of 0.85 and the 2.4% improvement claim in Section VI.A may be computed on training data or on a differently defined split, and cannot be interpreted as held-out test performance as claimed.
  3. [Section V.D.2, Table I] The paper states that 5-fold cross-validation is used and that paired t-tests and McNemar's test are applied, but no standard deviations, confidence intervals, or per-fold results are reported for any model or metric. Without variance estimates, the claimed statistical significance (e.g., chi-square = 8.34, p = 0.004) is unverifiable, and the superiority claim rests on point estimates alone, which is insufficient for a state-of-the-art empirical comparison.
  4. [Table II] The ablation results are internally inconsistent. 'Remove Multi-Head Attention' yields Macro F1 = 0.81, while 'Single Attention Head' yields Macro F1 = 0.84 with a 0.00 delta. Since removing multi-head attention should effectively produce a single-head model, these two configurations should give the same result; the discrepancy undermines the ablation-based attribution of performance to the multi-head mechanism.
minor comments (6)
  1. [Section V.D] Section V.D.1 describes a 70-15-15 train/validation/test split, while Section V.D.2 describes 5-fold cross-validation; the relationship between these two protocols is never clarified.
  2. [Section IV.B] The temporal encoding module is a two-layer MLP rather than a sequence model; the name 'Deep Temporal Networks' is misleading given the single-timestamp nature of the input.
  3. [Section IV.A, IV.E] Numerical values for several hyperparameters are never reported: the scaling parameter alpha in Eq. (5), the connectivity threshold lambda_min in Eq. (8), the connectivity-type weights phi in Eq. (6), and the focal-loss class weights alpha_c in Eq. (29). Without these values, the architecture is not reproducible from the text.
  4. [Section VII, Section VI.C] The paper claims that attention weights provide interpretability, but no analysis of learned attention weights is presented; none of the figures visualize or quantify attention patterns in relation to domain knowledge.
  5. [Table V] Table V labels the second column 'Training Time' but also reports inference time and memory in the same row; it should be clarified which phase each number refers to.
  6. [References] Several references appear incomplete or inconsistently formatted (e.g., [5] lacks a year and identifier, [25] lacks venue details); the reference list should be checked for completeness.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central claim is an empirical test-set comparison against external baselines, with no fitted quantity renamed as a prediction.

full rationale

The paper's derivation chain is an empirical machine-learning pipeline: it constructs a graph, defines features, trains a model with focal loss and AdamW, and reports held-out test metrics in Table I. No reported quantity is defined in terms of a fitted parameter that also appears as the output, and no equation in Sections III-VI makes a prediction equal to its input by construction. The ablation study (Table II) removes architectural components and remeasures performance; that is a standard sensitivity analysis, not a circular reduction. The literature references are external; there is no load-bearing self-citation chain, no imported uniqueness theorem, and no ansatz smuggled in under the authors' own prior authority. A data-validity concern does exist: Section V.A describes FARS as a record of fatal crashes, while Table IV reports 45.2% 'no injury' and 32.1% 'minor' classes and support counts summing to roughly the full dataset rather than the stated 15% test split. That inconsistency undermines the empirical claim's evidentiary basis, but it is a data-definition and evaluation-protocol problem, not a circular-reasoning pattern: the model's test predictions are not used to fit the model or to define the evaluation metric. Accordingly, the circularity score is 0.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central performance claims rest on free hyperparameters and data assumptions. The most consequential is the FARS severity-label assumption, which is internally contradicted. The model itself is a standard stack of known components with no new physical or structural entities.

free parameters (4)
  • alpha (k-NN scaling) = not reported
    Eq (5) sets adaptive neighborhood size k_adaptive = max(3, min(15, floor(rho_local * alpha))); alpha is described as determined empirically and is not given.
  • lambda_min (connectivity threshold) = 0.1
    Section IV.A requires the second smallest Laplacian eigenvalue to exceed 0.1 to validate graph connectivity; the threshold is chosen without independent justification.
  • connectivity type weights phi = topological 1.0, spatial 0.8, functional 0.6
    Eq (6) assigns hand-chosen weights to connectivity types; these affect edge weights and thus the graph structure.
  • class weights alpha_c in focal loss = not reported
    Eq (29) uses class-specific weights alpha_c to address imbalance; values are not stated and would typically be fitted to class frequencies.
assumptions (4)
  • domain assumption FARS contains all four severity classes with the distribution in Table IV.
    Sections V.A and Table IV; contradicted by the paper's own statement that FARS records only fatal crashes.
  • domain assumption DBSCAN clustering on GPS coordinates yields nodes that correspond to real road segments.
    Section IV.A; the clustering parameters (Eqs. 2-3) rely on road width and GPS error estimates that are not sourced.
  • domain assumption The feature set (elevation, slope, curvature, etc.) is available and accurately encoded from both datasets.
    Section IV.B; no preprocessing pipeline is provided to show how raw FARS and ARI fields become these features.
  • standard math The statistical tests (Friedman, Nemenyi, paired t, McNemar) are correctly applied to the multiclass results.
    Section VI.A-B; McNemar's test is for paired binary outcomes and the paper does not describe binarization, and p-values are inconsistent across tests.

how reviews work

0 comments
Cite this review

Pith. "Pith review of STARN-GAT: A Multi-Modal Spatio-Temporal Graph Attention Network for Accident Severity Prediction." pith.science (2026). https://pith.science/paper/7PWRRFV6

@misc{pith2026250720451,
  author       = {Pith},
  title        = {Pith review of: STARN-GAT: A Multi-Modal Spatio-Temporal Graph Attention Network for Accident Severity Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7PWRRFV6}},
  note         = {Machine review of arXiv:2507.20451}
}
read the original abstract

Accurate prediction of traffic accident severity is critical for improving road safety, optimizing emergency response strategies, and informing the design of safer transportation infrastructure. However, existing approaches often struggle to effectively model the intricate interdependencies among spatial, temporal, and contextual variables that govern accident outcomes. In this study, we introduce STARN-GAT, a Multi-Modal Spatio-Temporal Graph Attention Network, which leverages adaptive graph construction and modality-aware attention mechanisms to capture these complex relationships. Unlike conventional methods, STARN-GAT integrates road network topology, temporal traffic patterns, and environmental context within a unified attention-based framework. The model is evaluated on the Fatality Analysis Reporting System (FARS) dataset, achieving a Macro F1-score of 85 percent, ROC-AUC of 0.91, and recall of 81 percent for severe incidents. To ensure generalizability within the South Asian context, STARN-GAT is further validated on the ARI-BUET traffic accident dataset, where it attains a Macro F1-score of 0.84, recall of 0.78, and ROC-AUC of 0.89. These results demonstrate the model's effectiveness in identifying high-risk cases and its potential for deployment in real-time, safety-critical traffic management systems. Furthermore, the attention-based architecture enhances interpretability, offering insights into contributing factors and supporting trust in AI-assisted decision-making. Overall, STARN-GAT bridges the gap between advanced graph neural network techniques and practical applications in road safety analytics.

Figures

Figures reproduced from arXiv: 2507.20451 by the authors.

Figure 1
Figure 1. Graph construction methodology where 𝑎𝑘 ∈ 𝑅 𝟛𝑑 is the learned attention parameter vector for attention head 𝑘, || denotes vector concatenation, 𝑊𝑘 is the learned transformation matrix for head 𝑘, and 𝑒𝑖𝑗 represents edge features including normalized distance, road type similarity, and connectivity type indicators. Second, attention coefficients are normalized using the SoftMax function. α𝑖𝑗 (𝑘) = exp(𝑒𝑖𝑗 (𝑘) ) ∑ exp… view at source ↗
Figure 2
Figure 2. STARN-GAT Model Architecture Next, self-attention weights are computed using the scaled dot-product attention mechanism: 𝐴 = Softmax ( 𝑄𝑄 𝑇 √64) ∈ 𝑅 𝟛×𝟛 (22) The attended fusion is then computed as: 𝐻𝑓𝑢𝑠𝑒𝑑 = 𝐴𝑄 ∈ 𝑅 𝟛×𝟞𝟜 (23) Finally, the fused representation is flattened to create the input for the classification head: ℎ𝑓𝑖𝑛𝑎𝑙 = Flatten(𝐻𝑓𝑢𝑠𝑒𝑑) ∈ 𝑅 𝟙𝟡𝟚 (24) D. Classification Head and Output Layer Multi-Layer Classifi… view at source ↗
Figure 3
Figure 3. ROC curve of STARN-GAT model for both dataset [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5 [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Temporal pattern analysis [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 8
Figure 8. Figure 8: Computational performance analysis TABLE V. TABLE 5: DETAILED COMPUTATIONAL PERFORMANCE ANALYSIS System Training Time Memory Usage Inference Time Energy Consumption Scalability Factor GPU (V100) 3.2 hours 2.5 GB 125 ms 52 W·h 1.0× baseline GPU (RTX 3080) 4.1 hours 2.8 …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 52 canonical work pages

  1. [1]

    Predicting Severe Injury in Motor Vehicle Crashes,

    U.S. Department of Transportation, National Highway Traffic Safety Administration, "Predicting Severe Injury in Motor Vehicle Crashes," 2018

    work page 2018

  2. [2]

    Spatio-temporal analysis of road traffic crashes by severity,

    A. M. H. El-Basyouny and Y. Abdel-Aty, "Spatio-temporal analysis of road traffic crashes by severity," Accid. Anal. Prev., vol. 36, no. 5, pp. 845–853, Sep. 2004

    work page 2004

  3. [3]

    An overview of multinomial logistic regression for traffic accident severity classification,

    M. S. Abdulhafedh, "An overview of multinomial logistic regression for traffic accident severity classification," Int. J. Traffic Transp. Eng., vol. 4, no. 2, pp. 45–56, 2015

    work page 2015

  4. [4]

    Ordered probit models for motor vehicle crash injury severity in New Jersey,

    G. Ozbay, "Ordered probit models for motor vehicle crash injury severity in New Jersey," J. Transp. Eng., vol. 141, no. 1, p. 04014070, Jan. 2015

    work page 2015

  5. [5]

    Ensemble methods for accident severity prediction using regression, trees, and forests,

    A. Sam and A. Gulia, "Ensemble methods for accident severity prediction using regression, trees, and forests," SSRN, 2023

    work page 2023

  6. [6]

    Using support vector machine models for crash injury severity analysis,

    J. Yu and M. A. Abdel-Aty, "Using support vector machine models for crash injury severity analysis," in Proc. Transp. Res. Board 92nd Annu. Meeting, Washington, DC, USA, 2013, pp. 1–15

    work page 2013

  7. [7]

    Transportation network connectivity and accident analysis: A case study of Gainesville, Florida,

    B. Anderson and D. Hernandez, "Transportation network connectivity and accident analysis: A case study of Gainesville, Florida," J. Transp. Geogr., vol. 60, pp. 200–209, Apr. 2017

    work page 2017

  8. [8]

    Spatio -temporal graph convolutional networks: A comprehensive review,

    B. Yu, Y. Yin, and Z. Li, "Spatio -temporal graph convolutional networks: A comprehensive review," in Proc. Int. Joint Conf. Artif. Intell. (IJCAI), Stockholm, Sweden, Jul. 2018

    work page 2018

Show all 54 references
  1. [9]

    A survey on graph neural networks,

    S. Ma, D. Zhu, and Z. Fan, "A survey on graph neural networks," IEEE Trans. Knowl. Data Eng., vol. 33, no. 3, pp. 993–1008, Mar. 2021

    2021

  2. [10]

    Spectral networks and deep locally connected networks on graphs,

    J. Bruna, W. Zaremba, and Y. LeCun, "Spectral networks and deep locally connected networks on graphs," in Proc. Int. Conf. Learn. Represent. (ICLR), Banff, AB, Canada, Apr. 2014

    2014

  3. [11]

    Graph neural networks for road safety modeling: Datasets and evaluations for accident analysis,

    M. Li, J. Wu, and D. Hu, "Graph neural networks for road safety modeling: Datasets and evaluations for accident analysis," arXiv:2307.03058, 2023

    2023 arXiv

  4. [12]

    Multimodal fusion: Foundations, trends, and challenges,

    Y. Tian et al., "Multimodal fusion: Foundations, trends, and challenges," ACM Comput. Surv., vol. 55, no. 11, Art. no. 230, Nov. 2023, doi: 10.1145/3576920

    2023 doi

  5. [13]

    Cross-modal learning: Architectures and applications,

    L. Peng et al., "Cross-modal learning: Architectures and applications," IEEE Trans. Pattern Anal. Mach. Intell., vol. 46, no. 3, pp. 1237–1256, Mar. 2024, doi: 10.1109/TPAMI.2023.3331671

    2024

  6. [14]

    Spatio -temporal attention networks for traffic flow forecasting,

    X. Wang et al., "Spatio -temporal attention networks for traffic flow forecasting," IEEE Trans. Intell. Transp. Syst., vol. 24, no. 8, pp. 7890– 7903, Aug. 2023, doi: 10.1109/TITS.2023.3262114

    2023

  7. [15]

    DeepVision: Enhanced road safety prediction using multi-modal CNNs,

    R. Kumar and S. Liu, "DeepVision: Enhanced road safety prediction using multi-modal CNNs," arXiv:2401.04567, 2024

    2024 arXiv

  8. [16]

    A multi -modal attention neural network for traffic flow prediction by capturing long -short term sequence correlation,

    P. Tang, S. Yang, and Z. Li, "A multi -modal attention neural network for traffic flow prediction by capturing long -short term sequence correlation," J. Intell. Transp. Syst., vol. 26, no. 1, pp. 1–17, Jan. 2022

    2022

  9. [17]

    CNN-based models for accident severity prediction using road imagery,

    J. Park, Y. Kim, and S. Kim, "CNN-based models for accident severity prediction using road imagery," Accid. Anal. Prev., vol. 111, pp. 156– 167, Feb. 2018

    2018

  10. [18]

    LSTM networks for temporal dependencies in accident occurrence,

    Y. Zhang, Y. Wang, and M. Chen, "LSTM networks for temporal dependencies in accident occurrence," Transp. Res. Part C Emerg. Technol., vol. 92, pp. 297–310, Jul. 2018

    2018

  11. [19]

    Graph attention networks,

    P. Veličković et al., "Graph attention networks," in Proc. Int. Conf. Learn. Represent. (ICLR), Vancouver, BC, Canada, Apr. 2018

    2018

  12. [20]

    Graph convolutional networks for accident prediction,

    H. Guo, J. Li, and J. Liu, "Graph convolutional networks for accident prediction," IEEE Access, vol. 8, pp. 210087–210098, 2020

    2020

  13. [21]

    Accident hotspot identification using graph-based models,

    J. Jin, Y. Zhang, and H. Wang, "Accident hotspot identification using graph-based models," Transp. Res. Part B Methodol., vol. 129, pp. 1 – 15, Nov. 2019

    2019

  14. [22]

    Road network simplification preserving topology,

    C. Pung, W. B. Han, Y. A. Han, and K. Y. Cho, "Road network simplification preserving topology," Appl. Netw. Sci., vol. 7, no. 1, p. 19, 2022

    2022

  15. [23]

    A density -based algorithm for discovering clusters in large spatial databases with noise,

    M. Ester, H. -P. Kriegel, J. Sander, and X. Xu, "A density -based algorithm for discovering clusters in large spatial databases with noise," in Proc. 2nd Int. Conf. Knowl. Discov. Data Min. (KDD), Portland, OR, USA, Aug. 1996, pp. 226–231

    1996

  16. [24]

    The network analysis of urban streets: A dual approach,

    S. Porta, P. Crucitti, and V. Latora, "The network analysis of urban streets: A dual approach," Environ. Plan. B Plan. Des., vol. 33, no. 5, pp. 707–725, Oct. 2006

    2006

  17. [25]

    Applying social network analysis to road networks,

    J. Park and B. Yilmaz, "Applying social network analysis to road networks," in Proc. ASPRS Annu. Conf., San Diego, CA, USA, Apr. 2010, pp. 1–10

    2010

  18. [26]

    Algebraic connectivity of graphs,

    M. Fiedler, "Algebraic connectivity of graphs," Czech. Math. J., vol. 23, no. 2, pp. 298–305, 1973

    1973

  19. [27]

    Feature engineering for temporal data: Methodologies and industrial applications,

    M. Tsitsifli et al., "Feature engineering for temporal data: Methodologies and industrial applications," Eng. Appl. Artif. Intell., vol. 132, Art. no. 107882, Jun. 2024, doi: 10.1016/j.engappai.2024.107882

    2024

  20. [28]

    Inductive representation learning on large graphs,

    W. Hamilton, Z. Ying, and J. Leskovec, "Inductive representation learning on large graphs," in *Proc. 31st Int. Conf. Neural Inf. Process. Syst. (NeurIPS)*, Long Beach, CA, USA, 2017, pp. 1025–1035

    2017

  21. [29]

    Deep sparse rectifier neural networks,

    X. Glorot, A. Bordes, and Y. Bengio, "Deep sparse rectifier neural networks," in Proc. 14th Int. Conf. Artif. Intell. Statist. (AISTATS), Fort Lauderdale, FL, USA, Apr. 2011, pp. 315–323

    2011

  22. [30]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), Las Vegas, NV, USA, Jun. 2016, pp. 770–778

    2016

  23. [31]

    Layer normalization,

    J. L. Ba, J. R. Kiros, and G. E. Hinton, "Layer normalization," arXiv:1607.06450, 2016

    2016 arXiv

  24. [32]

    Batch normalization: Accelerating deep network training by reducing internal covariate shift,

    S. Ioffe and C. Szegedy, "Batch normalization: Accelerating deep network training by reducing internal covariate shift," in Proc. Int. Conf. Mach. Learn. (ICML), Lille, France, Jul. 2015, pp. 448–456

    2015

  25. [33]

    Attention is all you need,

    A. Vaswani et al., "Attention is all you need," in Adv. Neural Inf. Process. Syst. (NeurIPS), Long Beach, CA, USA, Dec. 2017, pp. 5998– 6008

    2017

  26. [34]

    Dropout: A simple way to prevent neural networks from overfitting,

    N. Srivastava et al., "Dropout: A simple way to prevent neural networks from overfitting," J. Mach. Learn. Res., vol. 15, pp. 1929 – 1958, Jun. 2014

    1929

  27. [35]

    Focal loss for dense object detection,

    T. Lin et al., "Focal loss for dense object detection," in Proc. IEEE Int. Conf. Comput. Vis. (ICCV), Venice, Italy, Oct. 2017, pp. 2980–2988

    2017

  28. [36]

    Decoupled weight decay regularization,

    I. Loshchilov and F. Hutter, "Decoupled weight decay regularization," in Proc. Int. Conf. Learn. Represent. (ICLR), New Orleans, LA, USA, May 2019

    2019

  29. [37]

    SGDR: Stochastic gradient descent with warm restarts,

    I. Loshchilov and F. Hutter, "SGDR: Stochastic gradient descent with warm restarts," in Proc. Int. Conf. Learn. Represent. (ICLR), Toulon, France, Apr. 2017

    2017

  30. [38]

    On the difficulty of training recurrent neural networks,

    R. Pascanu, T. Mikolov, and Y. Bengio, "On the difficulty of training recurrent neural networks," in Proc. Int. Conf. Mach. Learn. (ICML), Atlanta, GA, USA, Jun. 2013, pp. 1310–1318

    2013

  31. [39]

    Fatality Analysis Reporting System (FARS),

    U.S. Department of Transportation, National Highway Traffic Safety Administration, "Fatality Analysis Reporting System (FARS)," [Online]. Available: https://www.nhtsa.gov/research -data/fatality- analysis-reporting-system-fars. Accessed: May 28, 2024

    2024

  32. [40]

    Bangladesh Road Accident Database,

    ARI-BUET, "Bangladesh Road Accident Database," 2020. [Online]. Available: https://ari.buet.ac.bd/research/bradb/

    2020

  33. [41]

    Graph neural networks: A review of methods and applications,

    J. Zhou et al., "Graph neural networks: A review of methods and applications," AI Open, vol. 1, pp. 57–81, 2020

    2020

  34. [42]

    Spatial -temporal synchronous graph convolutional networks for traffic forecasting,

    K. Song, Z. Li, and J. Ma, "Spatial -temporal synchronous graph convolutional networks for traffic forecasting," in Proc. AAAI Conf. Artif. Intell., New York, NY, USA, Feb. 2020, pp. 6960–6967

    2020

  35. [43]

    ST -GraphNet: A spatio -temporal graph neural network for understanding and predicting automated vehicle crash severity,

    H. Zhang, Y. Wang, and J. Liu, "ST -GraphNet: A spatio -temporal graph neural network for understanding and predicting automated vehicle crash severity," arXiv:2403.04709, 2024

    2024 arXiv

  36. [44]

    A spatial-temporal graph gated transformer for traffic forecasting,

    H. Zhao, R. Ma, and Y. Li, "A spatial-temporal graph gated transformer for traffic forecasting," J. Adv. Transp., vol. 2020, Art. ID 8838323, 2020

    2020

  37. [45]

    ST-GTrans: Spatio-temporal graph transformers with semantic road embeddings,

    H. Zhang et al., "ST-GTrans: Spatio-temporal graph transformers with semantic road embeddings," arXiv:2402.18934, 2024

    2024 arXiv

  38. [46]

    Evaluation: From precision, recall and F -measure to ROC, informedness, markedness and correlation,

    D. M. Powers, "Evaluation: From precision, recall and F -measure to ROC, informedness, markedness and correlation," J. Mach. Learn. Technol., vol. 2, no. 1, pp. 37–63, Dec. 2011

    2011

  39. [47]

    A systematic analysis of performance measures for classification tasks,

    M. Sokolova and G. Lapalme, "A systematic analysis of performance measures for classification tasks," Inf. Process. Manag., vol. 45, no. 4, pp. 427–437, Jul. 2009

    2009

  40. [48]

    The balanced accuracy and its posterior distribution,

    K. H. Brodersen, C. S. Ong, K. E. Stephan, and J. M. Buhmann, "The balanced accuracy and its posterior distribution," in Proc. 20th Int. Conf. Pattern Recognit. (ICPR), Istanbul, Turkey, Aug. 2010, pp. 3121–3124

    2010

  41. [49]

    The precision -recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets,

    T. Saito and M. Rehmsmeier, "The precision -recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets," PLoS ONE, vol. 10, no. 3, Art. no. e0118432, Mar. 2015

    2015

  42. [50]

    An introduction to ROC analysis,

    T. Fawcett, "An introduction to ROC analysis," Pattern Recognit. Lett., vol. 27, no. 8, pp. 861–874, Jun. 2006

    2006

  43. [51]

    A coefficient of agreement for nominal scales,

    J. Cohen, "A coefficient of agreement for nominal scales," Educ. Psychol. Meas., vol. 20, no. 1, pp. 37–46, Apr. 1960

    1960

  44. [52]

    Public crash databases: Bias assessment and enhancement strategies,

    A. Richter et al., "Public crash databases: Bias assessment and enhancement strategies," Accid. Anal. Prev., vol. 198, Art. no. 107486, May 2024, doi: 10.1016/j.aap.2024.107486

    2024

  45. [53]

    A study of cross -validation and bootstrap for accuracy estimation and model selection,

    R. Kohavi, "A study of cross -validation and bootstrap for accuracy estimation and model selection," in Proc. Int. Joint Conf. Artif. Intell. (IJCAI), Montreal, QC, Canada, Aug. 1995, pp. 1137–1143

    1995

  46. [54]

    Multiple comparisons among means,

    O. J. Dunn, "Multiple comparisons among means," J. Amer. Statist. Assoc., vol. 56, no. 293, pp. 52–64, Mar. 1961

    1961

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.