Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Towards Reliable and Interpretable Traffic Crash Pattern Prediction and Safety Interventions Using Customized Large Language Models

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Fine-tuned LLMs beat tabular models at predicting crash severity, injury counts, and crash type, by reasoning over crash reports turned into text.

desk verdict This paper is worth a look for its dataset, but the headline 41.7% F1 gain is not yet supported by the evaluation as described. read the letter →

arxiv 2505.12545 v2 pith:V763373B submitted 2025-05-18 cs.CL

Yang Zhao , Pu Wang , Yibo Zhao , Hongru Du , Hao Frank Yang This is my paper · ORCID
classification cs.CL
keywords trafficcrashpredictionlargelanguagemodelsfine-tuningseverityfeatureattributionShapleyvaluesconditionalriskanalysistext-basedreasoning
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

This paper tries to show that reframing traffic-crash outcome prediction as a text-reasoning task—rather than feeding tabular features into classical machine-learning classifiers—produces substantially more accurate and more interpretable predictions. The authors build a large textual dataset from real crash records, fine-tune Llama 3.1 models to predict injury count, severity, and crash type, and report an average 41.7% F1-score improvement over seven tabular baselines. They also claim the model's confidence score is a calibrated risk proxy, reaching 97.61% precision for fatal-crash predictions when confidence exceeds 0.6. If true, this would give traffic-safety agencies a single pipeline that both forecasts crash outcomes and attributes risk to specific contributing factors, supporting targeted interventions.

What carries the argument

The load-bearing machinery is the TrafficSafe Event dataset plus the supervised fine-tuning of Llama 3.1 with special output tokens. Raw multi-modal crash data (crash reports, infrastructure, vehicle, person, and satellite-image-derived text) is converted through an AI-expert cooperative textualization process into natural-language prompts. The prediction targets are added to the model's vocabulary as special tokens, and the model is fine-tuned with LoRA to generate these tokens, turning classification into next-token generation. The second key mechanism is TrafficSafe Attribution, which computes Shapley values over sentence-level feature groups to assign contributions that sum to the model's confidence score, enabling conditional risk analysis.

What would settle it

A direct check: retrain the seven tabular baselines on the original feature set, then evaluate them on the exact same rebalanced test sets that TrafficSafe uses, after adding the satellite-image-derived and textual features that the prompts contain. If the baselines' F1-scores rise to near the LLM's levels, the 41.7% gain is attributable to added information or evaluation-set changes rather than to LLM text reasoning.

Watch

Extended reading notes

Core claim

The central claim is that a language model fine-tuned on textualized multi-modal crash records (the TrafficSafe Event dataset, 58,903 prompts, about 12.74 million words) outperforms standard tabular machine-learning baselines across three crash-outcome prediction tasks: number of injuries, injury severity, and crash type. The paper reports an average F1-score improvement of 41.7% over the baselines, with the largest gains on crash-type prediction (e.g., F1 of 0.759 in Washington versus at most 0.329 for the best baseline). The paper further claims that the model's confidence score is a reliable indicator of risk: accuracy rises with confidence, and for fatal-crash predictions precision reaches 97.61% when confidence exceeds 0.6. The paper also introduces TrafficSafe Attribution, a Shapley-based sentence-level attribution method, and uses it to identify alcohol-impaired driving as the leading contributor to severe crashes, with work-zone plus alcohol combinations and multi-factor co-occurrence markedly raising predicted risk.

Load-bearing premise

The headline improvement over tabular baselines assumes that those baselines were given the same information content and the same rebalanced evaluation sets as the LLM, so the gain comes from text reasoning rather than from extra text or satellite data or from the evaluation-set rebalancing.

Editorial extensions

If this is right

  • Traffic agencies could deploy a single fine-tuned LLM to predict injury counts, severity, and crash type from text-based incident descriptions, replacing or complementing tabular classifiers.
  • The confidence score could be used as a decision threshold: a fatal-crash prediction with confidence above 0.6 corresponds to 97.61% precision, enabling prioritization of high-risk cases for review or intervention.
  • Sentence-level attribution could be used to run what-if risk analyses for rare or unseen conditions (e.g., hypothetical alcohol-plus-work-zone scenarios), guiding policy changes even where crash data is sparse.
  • Training-stage attribution identifies which data components (unit information, event information) most improve model performance, providing a data-collection priority list for states building crash datasets.
  • The framework's zero-shot generalization results (F1 around 0.52-0.54 on unseen North Carolina and Maine data) suggest the model captures transferable crash patterns across states.

Reading between the lines

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

  • The 41.7% F1 improvement may partially reflect that the baselines were evaluated with class-imbalanced data, while the LLM's test sets were rebalanced by removing 1,428 S1 (no-apparent-injury) records for Washington; a fair apples-to-apples comparison would rerun baselines on the identical rebalanced test sets with the same information content.
  • A testable extension would be to probe whether the Shapley-based attributions track the causal contributions of factors (e.g., alcohol, work zones) or merely the statistical associations learned from the training corpus; controlled counterfactual generations with edited prompts could separate these.
  • The confidence-precision calibration claim suggests a promising practical safeguard: refusal-to-predict or deferral-to-human-review rules based on confidence thresholds could be embedded in real-time crash-response or safety-policy workflows, and could be validated prospectively on new crash years.
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 / 5 minor

Summary. The paper introduces TrafficSafe, a framework that textualizes multi-modal traffic crash data (HSIS records, satellite images, driver and vehicle information) into natural-language prompts and fine-tunes Llama 3.1 (8B and 70B) with LoRA and task-specific special tokens for three prediction targets: number of injuries, injury severity, and crash type. The authors report that TrafficSafe achieves a 41.7% average F1-score improvement over seven tabular baselines (Section 3.3, Table 1), that higher confidence scores correlate with higher accuracy and that fatal-crash precision reaches 97.61% when confidence exceeds 0.6 (Figures 3e-3i), and that a sentence-level Shapley attribution method (TrafficSafe Attribution) supports conditional risk analysis, with findings such as alcohol-impaired driving and work-zone combinations being highest-risk. The paper also reports zero-shot generalization to North Carolina and Maine and provides code and processed data examples via GitHub.

Significance. The paper addresses a timely and societally relevant problem, and the idea of reframing crash prediction as a text-based reasoning task with interpretable attributions is a potentially valuable contribution to traffic safety analytics. The dataset construction (58,903 records, 12.74 million words), the use of real-world HSIS data, and the public availability of code and data examples are concrete assets for reproducibility. If the comparative claims and the confidence-as-risk calibration are established after the evaluation confounds are addressed, the framework could offer a meaningful advance over conventional tabular crash prediction. The attribution analysis, if carefully framed, could also be useful for guiding data collection. However, the central quantitative claims currently rest on an evaluation protocol whose controls are not fully documented.

major comments (4)
  1. [Section 5.2.4 and Table 1] The reported F1 improvements are measured on test sets that were rebalanced by removing 1,428 S1 records in the Washington test/validation split and by balancing all subsets in Illinois, while Section 5.2.6 does not report any equivalent class-imbalance handling for the tabular baselines. Since TrafficSafe was trained with uniform class sampling (Section 5.2.3) and the baselines are not described as using class weights or resampling, the comparison is not controlled for class-imbalance handling; the paper should report results on the original unmodified held-out test distribution as well, or specify and justify the same rebalancing procedure for all methods.
  2. [Section 5.2.6 and Section 5.1.2] The baseline feature sets are never specified. The TrafficSafe prompts contain GPT-4o-derived satellite-image annotations and narrative descriptions of driver behavior, vehicle movement, and occupant details, but the paper does not state which subset of raw HSIS fields the RandomForest, XGBoost, CatBoost, and other baselines received as input. Without this information, the 41.7% F1 gain cannot be attributed to text-based reasoning rather than to the inclusion of additional data modalities; the authors should document the exact features used for each baseline and, ideally, add a text-based baseline (e.g., TF-IDF with a linear classifier trained on the same textualized prompts) to isolate the contribution of the LLM.
  3. [Section 4 and Figure 3i] The claimed 97.61% precision for fatal predictions when confidence exceeds 0.6, and the related statement in the Discussion that this reflects 'real-world occurrence risk', are computed on the rebalanced test sets described in Section 5.2.4, where 1,428 majority-class S1 records were removed from the Washington split and the Illinois subsets were balanced. Because the test-set class priors were intentionally altered, this precision value is not a population-level risk estimate; the paper should either recompute the confidence-calibration analysis on an unmodified representative test set or explicitly state that the number is conditional on the rebalanced evaluation distribution.
  4. [Section 3.4 and Abstract] The conditional-risk findings, such as 'alcohol-impaired driving is the leading factor in severe crashes' and 'aggressive and impairment-related behaviors have nearly twice the contribution for severe crashes', are derived from the model's confidence score used as a proxy for risk and from Shapley attributions of that confidence (Equations 7-8). These are statements about the model's internal attribution and have not been validated as population causal risk factors; the abstract and Section 3.4 should be rephrased to make clear that these are model-attributed contributions to predicted confidence rather than established real-world risk factors, or the confidence-as-risk assumption should be validated on held-out outcome data.
minor comments (5)
  1. [Section 5.2.5] The metrics in Equations (2)-(4) are defined for binary classification, but the tasks (Injury, Severity, Type) are multi-class; the paper should specify how precision and F1 are averaged across classes and whether the reported numbers are macro-, micro-, or class-weighted.
  2. [Section 3.3] No confidence intervals, error bars, or statistical significance tests are reported for the F1 differences in Table 1; reporting a bootstrap or repeated-seed analysis would strengthen the comparison, especially for the rare fatal and serious-injury classes.
  3. [Section 5.3.3] Equation (8) defines the inference-stage Shapley value using the LLM's predicted probability, but the text does not specify which class probability is used (e.g., the probability of the predicted special token) or how the stratified sampling estimator of Reference 49 is configured; a brief pseudocode or explicit estimator formula would improve reproducibility.
  4. [Section 5.1.2] The GPT-4o textualization process (satellite-image annotation and prompt generation) should explicitly state that this step is done before the train/test split and uses no label information, to reassure readers that the test prompts are not influenced by the training labels.
  5. [Extended Data Figure 3 and Figure 3c/3d] The confusion matrices show that baselines collapse onto the majority classes; reporting classwise F1-scores for the rare classes (e.g., 'serious injury' and 'fatal') in the main text would help readers assess whether the TrafficSafe improvement is driven mainly by rare-class recall.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the main F1 claim is anchored by held-out labels, and the attribution statements are model interpretations rather than predictions derived from their own inputs.

full rationale

The central load-bearing claim—a 41.7% average F1 improvement over tabular baselines—is evaluated against held-out labels in Table 1 using the 7:1.5:1.5 split described in Section 5.2.4; the labels are external to the model and are not constructed from the model's outputs, so the comparison is not circular. The confidence-as-risk analysis is an explicit modeling choice (Section 3.4: 'we can use the confidence score to represent a case's real-world risk level'), and the statement that the sum of feature contributions equals the confidence score is a Shapley efficiency property of Eq. 8, not a prediction derived from itself. The factor-attribution findings are internal interpretations of the fitted model rather than fitted parameters renamed as predictions. The only self-citation, Ref. 50 for uniform sampling in Section 5.2.3, is not load-bearing. Concerns that baselines may have used different feature sets or different class-imbalance handling affect the validity and generalizability of the comparison, but they are experimental confounds, not circular derivations. No uniqueness theorem or author-imported ansatz is invoked. Therefore no circular step is established.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

This is an empirical ML paper; the central claim rests on data quality and evaluation protocol assumptions rather than on analytical first principles. The free parameters listed are evaluation choices that influence the headline numbers, and the axioms capture the unverified premises in the data construction and risk interpretation.

free parameters (3)
  • Confidence threshold 0.6 = 0.6
    Used to claim 97.61% precision for fatal predictions in Figure 3i; chosen post hoc on test data.
  • Washington test-set rebalancing (removal of S1 cases) = 1,428 records removed
    Validation and test sets were balanced by removing 1,428 no-apparent-injury records (Section 5.2.4), altering the evaluation distribution.
  • Shapley stratified sampling configuration = not specified
    Equation (8) relies on a stratified sampling approximation whose sample size and variance are not reported.
assumptions (4)
  • domain assumption GPT-4o textualization preserves the information in the raw structured data without introducing systematic hallucination
    The entire TrafficSafe Event Dataset is built by translating key-value pairs into paragraphs with GPT-4o (Section 5.1.2); no quality check is reported.
  • ad hoc to paper The rebalanced Washington validation/test sets (S1 removed) support valid model comparison
    Section 5.2.4 removes most no-apparent-injury cases to balance evaluation; this changes the class distribution and may favor models with balanced training.
  • domain assumption LLM confidence scores are a valid proxy for real-world crash risk
    Section 3.4 uses confidence as risk level and draws policy conclusions; calibration is shown only on the rebalanced test set.
  • standard math Stratified sampling estimates Shapley values accurately
    Equation (8) in Section 5.3.3 uses a sampling-based Shapley estimator; the approximation error is not quantified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Reliable and Interpretable Traffic Crash Pattern Prediction and Safety Interventions Using Customized Large Language Models." pith.science (2026). https://pith.science/paper/V763373B

@misc{pith2026250512545,
  author       = {Pith},
  title        = {Pith review of: Towards Reliable and Interpretable Traffic Crash Pattern Prediction and Safety Interventions Using Customized Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/V763373B}},
  note         = {Machine review of arXiv:2505.12545}
}
read the original abstract

Predicting crash events is crucial for understanding crash distributions and their contributing factors, thereby enabling the design of proactive traffic safety policy interventions. However, existing methods struggle to interpret the complex interplay among various sources of traffic crash data, including numeric characteristics, textual reports, crash imagery, environmental conditions, and driver behavior records. As a result, they often fail to capture the rich semantic information and intricate interrelationships embedded in these diverse data sources, limiting their ability to identify critical crash risk factors. In this research, we propose TrafficSafe, a framework that adapts LLMs to reframe crash prediction and feature attribution as text-based reasoning. A multi-modal crash dataset including 58,903 real-world reports together with belonged infrastructure, environmental, driver, and vehicle information is collected and textualized into TrafficSafe Event Dataset. By customizing and fine-tuning LLMs on this dataset, the TrafficSafe LLM achieves a 42% average improvement in F1-score over baselines. To interpret these predictions and uncover contributing factors, we introduce TrafficSafe Attribution, a sentence-level feature attribution framework enabling conditional risk analysis. Findings show that alcohol-impaired driving is the leading factor in severe crashes, with aggressive and impairment-related behaviors having nearly twice the contribution for severe crashes compared to other driver behaviors. Furthermore, TrafficSafe Attribution highlights pivotal features during model training, guiding strategic crash data collection for iterative performance improvements. The proposed TrafficSafe offers a transformative leap in traffic safety research, providing a blueprint for translating advanced AI technologies into responsible, actionable, and life-saving outcomes.

Figures

Figures reproduced from arXiv: 2505.12545 by the authors.

Figure 1
Figure 1. Overview of the Proposed TrafficSafe Framework. (a) The U.S. faces one of the highest crash risks among developed countries, with a rising trend. However, analyzing and addressing this issue is challenging due to the heterogeneous factors involved in crash events, including traffic conditions, human behavior, environmental impacts, and driver characteristics. To tackle this, we propose TrafficSafe, a framework desig… view at source ↗
Figure 2
Figure 2. TrafficSafe Crash Outcomes Prediction Pipeline. Multi-modal crash data is collected and organized into textual prompts through an AI-expert cooperative process. The HSIS crash data, satellite images, and infrastructure data are used to extract general and infrastructure information, including the crash time, location, the road level, and so on. The vehicle data, and person data are converted into the event informati… view at source ↗
Figure 3
Figure 3. TrafficSafe LLM Provides Accurate and Trustworthy Predictions. TrafficSafe LLM produces robust confusion matrices for both the (a) Washington and (b) Illinois datasets (we select the best results for each task by F1-score). In contrast, baseline models tend to predict the most frequent category across both the (c) Washington and (d) Illinois datasets (we show baseline models with the best F1-score. The performances … view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Single Case Feature Attribution Results for Severity Task. The left part displays the full prompt from (a) Washington and (b) Illinois, with different colors representing various semantic text sequences. The right part illustrates the feature contribution assigned to e…
Figure 5
Figure 5. Figure 5: Conditional Risk Analysis for the Serious Injury and Fatal Crashes. Higher confidence scores in TrafficSafe LLM’s predictions correspond to greater accuracy, allowing the confidence score (calculated as the sum of feature contributions for all data components) to serve…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Benchmarking Frontier Large Language Models Against Official Crash Database Coding Using Police Crash Narratives

    cs.LG 2026-07 conditional novelty 6.0 of 10

    On 4,194 matched Arkansas fatal crashes, frontier LLMs did not beat a crude keyword-rule baseline at reproducing official crash coding, and agreement varied more by crash attribute than by model.

Reference graph

Works this paper leans on

57 extracted references · 50 canonical work pages · cited by 1 Pith paper

  1. [1]

    Vision Zero Action Plans Accessed: 2025-01-15

    Federal Highway Administration (FHW A). Vision Zero Action Plans Accessed: 2025-01-15. 2025

    work page 2025

  2. [2]

    Road Safety Annual Report 2023 (OECD Publishing, Paris, 2023)

    International Transport Forum (ITF). Road Safety Annual Report 2023 (OECD Publishing, Paris, 2023)

    work page 2023

  3. [3]

    R., Wang, D

    Islam, M. R., Wang, D. & Abdel-Aty, M. Calibrated confidence learning for large-scale real-time crash and severity prediction. npj Sustainable Mobility and Transport 1, 1 (2024)

    work page 2024

  4. [4]

    & Taniform, P

    Bougna, T., Hundal, G. & Taniform, P. Quantitative analysis of the social costs of road traffic crashes literature. Accident Analysis & Prevention 165, 106282 (2022)

    work page 2022

  5. [5]

    Wen, X., Xie, Y ., Jiang, L., Pu, Z. & Ge, T. Applications of machine learning methods in traffic crash severity modelling: current status and future directions. Transport reviews41, 855–879 (2021)

    work page 2021

  6. [6]

    R., Shankar, V

    Mannering, F., Bhat, C. R., Shankar, V . & Abdel-Aty, M. Big data, traditional data and the tradeoffs between prediction and causality in highway-safety analysis. Analytic methods in accident research25, 100113 (2020)

    work page 2020

  7. [7]

    & Liu, H

    Yan, X., Zou, Z., Feng, S., Zhu, H., Sun, H. & Liu, H. X. Learning naturalistic driving environment with statistical realism. Nature communications 14, 2037 (2023)

    work page 2023

  8. [8]

    & Grekul, J

    Tazul Islam, M., Thue, L. & Grekul, J. Understanding traffic safety culture: implications for increasing traffic safety. Transportation Research Record2635, 79–89 (2017)

    work page 2017

Show all 57 references
  1. [9]

    Mannering, F. L. & Bhat, C. R. Analytic methods in accident research: Methodological frontier and future directions. Analytic methods in accident research 1, 1–22 (2014)

    2014

  2. [10]

    & Xiong, Z

    Dong, C., Shao, C., Li, J. & Xiong, Z. An improved deep learning model for traffic crash prediction. Journal of Advanced Transportation 2018, 3869106 (2018)

    2018

  3. [11]

    & Antoniou, C

    Theofilatos, A., Chen, C. & Antoniou, C. Comparing machine learning and deep learning methods for real-time crash prediction. Transportation research record2673, 169–178 (2019). 16/35

    2019

  4. [12]

    Rahim, M. A. & Hassan, H. M. A deep learning based traffic crash severity prediction framework. Accident Analysis & Prevention 154, 106090 (2021)

    2021

  5. [13]

    & Masiur Rahman, S

    Sattar, K., Chikh Oughali, F., Assi, K., Ratrout, N., Jamal, A. & Masiur Rahman, S. Transparent deep machine learning framework for predicting traffic crash severity. Neural Computing and Applications 35, 1535–1547 (2023)

    2023

  6. [14]

    Data-driven risk analysis of nonlinear factor interactions in road safety using Bayesian networks

    Carrodano, C. Data-driven risk analysis of nonlinear factor interactions in road safety using Bayesian networks. Scientific Reports 14, 18948 (2024)

    2024

  7. [15]

    & Haque, M

    Sharma, A., Zheng, Z., Kim, J., Bhaskar, A. & Haque, M. M. Is an informed driver a better decision maker? A grouped random parameters with heterogeneity-in-means approach to investigate the impact of the connected environment on driving behaviour in safety-critical situations....

    2020

  8. [16]

    Xu, C., Ding, Z., Wang, C. & Li, Z. Statistical analysis of the patterns and characteristics of connected and autonomous vehicle involved crashes. Journal of safety research 71, 41–47 (2019)

    2019

  9. [17]

    & Ding, S

    Abdel-Aty, M. & Ding, S. A matched case-control analysis of autonomous vs human-driven vehicle accidents. Nature Communications 15, 4931 (2024)

    2024

  10. [18]

    M., Wali, B

    Boggs, A. M., Wali, B. & Khattak, A. J. Exploratory analysis of automated vehicle crashes in California: A text analytics & hierarchical Bayesian heterogeneity-based approach. Accident Analysis & Prevention 135, 105354 (2020)

    2020

  11. [19]

    Wali, B., Khattak, A. J. & Karnowski, T. The relationship between driving volatility in time to collision and crash-injury severity in a naturalistic driving environment. Analytic methods in accident research 28, 100136 (2020)

    2020

  12. [20]

    S., Cohen, J

    Ahmed, S. S., Cohen, J. & Anastasopoulos, P. C. A correlated random parameters with heterogeneity in means approach of deer-vehicle collisions and resulting injury-severities. Analytic methods in accident research 30, 100160 (2021)

    2021

  13. [21]

    & Kong, X

    Liu, Z., Chen, Y ., Xia, F., Bian, J., Zhu, B., Shen, G. & Kong, X. Tap: Traffic accident profiling via multi-task spatio-temporal graph representation learning. ACM Transactions on Knowledge Discovery from Data17, 1–25 (2023)

    2023

  14. [22]

    & Hansen, M

    Lu, J., Grembek, O. & Hansen, M. Learning the representation of surrogate safety measures to identify traffic conflict. Accident Analysis & Prevention 174, 106755 (2022)

    2022

  15. [23]

    L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., et al

    Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F. L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774 (2023)

    2023 arXiv

  16. [24]

    The llama 3 herd of models

    Meta, A., Jauhri, A., Pandey, A., Kadian, A., Al-Dahle, A., Letman, A., Mathur, A., Schelten, A., Yang, A., Fan, A., et al. The llama 3 herd of models. arXiv preprint arXiv:2407.21783 2 (2024)

    2024 arXiv

  17. [25]

    Gao, C., Lan, X., Lu, Z., Mao, J., Piao, J., Wang, H., Jin, D. & Li, Y . S3: Social-network Simulation System with Large Language Model-Empowered Agents. arXiv preprint arXiv:2307.14984 (2023). 17/35

    2023 arXiv

  18. [27]

    & Calafate, C

    De Zarzà, I., de Curtò, J., Roig, G. & Calafate, C. T. LLM multimodal traffic accident forecasting.Sensors 23, 9225 (2023)

    2023

  19. [28]

    Zhen, H., Shi, Y ., Huang, Y ., Yang, J. J. & Liu, N. Leveraging Large Language Models with Chain-of-Thought and Prompt Engineering for Traffic Crash Severity Analysis and Inference. Computers 13, 232 (2024)

    2024

  20. [29]

    & Ksaibati, K

    Rezapour, M. & Ksaibati, K. Identification of factors associated with various types of impaired driving. Humanities and social sciences communications 9, 1–11 (2022)

    2022

  21. [30]

    Department of Transportation, Federal Highway Administration

    U.S. Department of Transportation, Federal Highway Administration. Highway Safety Information System (HSIS) https://highways.dot.gov. Accessed: January 13, 2025. 2025

    2025

  22. [31]

    Google Maps Static API Documentation https://developers.google.com/maps/ documentation/maps-static

    Developers, G. Google Maps Static API Documentation https://developers.google.com/maps/ documentation/maps-static. Accessed: January 13, 2025. 2025

    2025

  23. [32]

    J., Shen, Y ., Wallis, P., Allen-Zhu, Z., Li, Y ., Wang, S., Wang, L

    Hu, E. J., Shen, Y ., Wallis, P., Allen-Zhu, Z., Li, Y ., Wang, S., Wang, L. & Chen, W. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685 (2021)

    2021 arXiv

  24. [33]

    & Brady, P

    Abdel-Aty, M., Keller, J. & Brady, P. A. Analysis of types of crashes at signalized intersections by using complete crash data and tree-based regression. Transportation Research Record1908, 37–45 (2005)

    2005

  25. [34]

    & Khattak, A

    Iranitalab, A. & Khattak, A. Comparison of four statistical and machine learning methods for crash severity prediction. Accident Analysis & Prevention 108, 27–36. ISSN : 0001-4575 (2017)

    2017

  26. [35]

    T., Mannering, F

    Savolainen, P. T., Mannering, F. L., Lord, D. & Quddus, M. A. The statistical analysis of highway crash-injury severities: A review and assessment of methodological alternatives. Accident Analysis & Prevention 43, 1666–1676. ISSN : 0001-4575 (2011)

    2011

  27. [36]

    Random Forests

    Breiman, L. Random Forests. Machine Learning 45, 5–32 (Oct. 2001)

    2001

  28. [37]

    & Schapire, R

    Freund, Y . & Schapire, R. E. A Decision-Theoretic Generalization of On-Line Learning and an Application to Boosting. Journal of Computer and System Sciences 55, 119–139. ISSN : 0022-0000 (1997)

    1997

  29. [38]

    Prokhorenkova, L., Gusev, G., V orobev, A., Dorogush, A. V . & Gulin, A.CatBoost: unbiased boosting with categorical features 2019. arXiv: 1706.09516 [cs.LG]

    2019 arXiv

  30. [39]

    Quinlan, J. R. Induction of decision trees. Machine learning 1, 81–106 (1986)

    1986

  31. [40]

    Cox, D. R. The regression analysis of binary sequences. Journal of the Royal Statistical Society Series B: Statistical Methodology 20, 215–232 (1958)

    1958

  32. [41]

    Chen, T. & Guestrin, C.XGBoost: A Scalable Tree Boosting Systemin Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (Association for Computing Machinery, San Francisco, California, USA, 2016), 785–794. ISBN : 9781450342322

    2016

  33. [42]

    S., Khazraee, S

    Lord, D., Geedipally, S., Pratt, M., Park, E. S., Khazraee, S. & Fitzpatrick, K. Safety Prediction Models for Six-Lane and One-Way Urban and Suburban Arterials ISBN : 978-0-309-29560-4 (Mar. 2022)

    2022

  34. [43]

    & von Luxburg, U

    Bordt, S. & von Luxburg, U. From shapley values to generalized additive models and backin International Conference on Artificial Intelligence and Statistics (2023), 709–745. 18/35

    2023

  35. [44]

    Shapley, L. S. in Contributions to the Theory of Games II (eds Kuhn, H. W. & Tucker, A. W.) 307–317 (Princeton University Press, Princeton, 1953)

    1953

  36. [45]

    Revised Code of Washington: Driving under the influencehttps://app

    Washington State Legislature. Revised Code of Washington: Driving under the influencehttps://app. leg.wa.gov/rcw/default.aspx?cite=46.61.502. Accessed: 2025-01-19. 2025

    2025

  37. [46]

    Accessed: 2025-01-19

    Illinois Secretary of State.Driving Under the Influence (DUI)https://www.ilsos.gov/departments/ drivers/traffic_safety/DUI/home.html. Accessed: 2025-01-19. 2025

    2025

  38. [47]

    M., Sourav, M

    Bhuiyan, H., Ara, J., Hasib, K. M., Sourav, M. I. H., Karim, F. B., Sik-Lanyi, C., Governatori, G., Rakotoni- rainy, A. & Yasmin, S. Crash severity analysis and risk factors identification based on an alternate data source: a case study of developing country. Scientific report...

    2022

  39. [48]

    Zhang, J., Huang, J., Jin, S. & Lu, S. Vision-language models for vision tasks: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence (2024)

    2024

  40. [49]

    pedalcyclist struck by vehicle

    Zhang, J., Sun, Q., Liu, J., Xiong, L., Pei, J. & Ren, K. Efficient Sampling Approaches to Shapley Value Approximation. Proc. ACM Manag. Data 1 (May 2023). 19/35 5 Methods 5.1 TrafficSafe Event Dataset Construction As introduced in Section 3.1, the raw crash data is multi-moda...

    2023

  41. [50]

    Du, H., Zhao, J., Zhao, Y ., Xu, S., Lin, X., Chen, Y ., Gardner, L. M. & Yang, H. F. Advancing Real- time Pandemic Forecasting Using Large Language Models: A COVID-19 Case Study. arXiv preprint arXiv:2404.06962 (2024)

    2024 arXiv

  42. [51]

    A., Ray, S

    Ahmed, S., Hossain, M. A., Ray, S. K., Bhuiyan, M. M. I. & Sabuj, S. R. A study on road accident prediction and contributing factors using explainable machine learning models: Analysis and performance. Transportation research interdisciplinary perspectives19, 100814 (2023)

    2023

  43. [52]

    & Abe, N

    Freund, Y ., Schapire, R. & Abe, N. A short introduction to boosting.Journal-Japanese Society For Artificial Intelligence 14, 1612 (1999)

    1999

  44. [53]

    Prokhorenkova, L., Gusev, G., V orobev, A., Dorogush, A. V . & Gulin, A. CatBoost: unbiased boosting with categorical features. Advances in neural information processing systems 31 (2018). 26/35

    2018

  45. [54]

    Chen, H., Lundberg, S. M. & Lee, S.-I. Explaining a series of models by propagating Shapley values. Nature communications 13, 4512 (2022)

    2022

  46. [55]

    C., Lundberg, S

    Chen, H., Covert, I. C., Lundberg, S. M. & Lee, S.-I. Algorithms to estimate Shapley value feature attributions. Nature Machine Intelligence 5, 590–601 (2023)

    2023

  47. [56]

    Lundberg, S. M. & Lee, S. -I. A Unified Approach to Interpreting Model Predictions in Advances in Neu- ral Information Processing Systems (eds Guyon, I., Luxburg, U. V ., Bengio, S., Wallach, H., Fergus, R., Vishwanathan, S. & Garnett, R.) 30 (Curran Associates, Inc., 2017)

    2017

  48. [57]

    & Horovicz, M

    Goldshmidt, R. & Horovicz, M. TokenSHAP: Interpreting Large Language Models with Monte Carlo Shapley Value Estimation. arXiv preprint arXiv:2407.10114 (2024)

    2024 arXiv

  49. [58]

    & Robnik-Šikonja, M

    Kokalj, E., Škrlj, B., Lavraˇc, N., Pollak, S. & Robnik-Šikonja, M. BERT meets Shapley: Extending SHAP Explanations to Transformer-based Classifiers in Proceedings of the EACL Hackashop on News Media Content Analysis and Automated Report Generation (eds Toivonen, H. & Boggia, ...

    2021

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.