{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,17]],"date-time":"2026-10-17T05:38:15Z","timestamp":1792215495637,"version":"4.4.0"},"reference-count":37,"publisher":"Public Library of Science (PLoS)","issue":"4","license":[{"start":{"date-parts":[[2026,4,28]],"date-time":"2026-04-28T00:00:00Z","timestamp":1777334400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:p>Road traffic crashes pose a serious public safety challenge, particularly due to fatal and serious injuries. Although machine learning (ML) has been widely used for crash severity prediction, many studies remain accuracy-oriented and insufficiently address class imbalance, decision thresholds, and probabilistic reliability. This study proposes a safety-oriented and explainable ML framework for predicting killed or seriously injured (KSI) crashes using nationwide United Kingdom traffic accident data from 2020\u20132024. Crash severity is reformulated as a binary classification task distinguishing slight injury crashes from KSI outcomes, aligning model objectives with road safety priorities. A Light Gradient Boosting Machine (LightGBM) model is developed with imbalance handling using SMOTE, safety-oriented decision threshold optimization, and probability calibration. Model performance is evaluated using ROC\u2013AUC, precision\u2013recall analysis, confusion matrices, the Brier score, and a utility-based evaluation metric, while interpretability is ensured through SHapley Additive exPlanations (SHAP). Results show that default threshold settings fail to adequately detect severe crashes. At an optimized threshold of 0.35, the model achieves a Recall(KSI) of 0.605 \u2013 representing a substantial 73% improvement compared to conventional configurations \u2013 while maintaining acceptable precision. In addition, probability calibration confirms reliable risk estimation (Brier score\u2009=\u20090.190), supporting risk-based interpretation. Comparative analysis demonstrates that the SMOTE-based model provides a more balanced and operationally effective trade-off than class-weighted learning. SHAP analysis identifies speed limit, road class, lighting conditions, and urban context as key variables associated with KSI risk. The findings highlight the importance of safety-oriented learning design and context-aware performance interpretation for effective, risk-based traffic safety management.<\/jats:p>","DOI":"10.1371\/journal.pone.0347873","type":"journal-article","created":{"date-parts":[[2026,4,28]],"date-time":"2026-04-28T17:48:25Z","timestamp":1777398505000},"page":"e0347873","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":3,"title":["Safety-oriented and explainable machine learning for KSI crash risk prediction: Evidence from the United Kingdom"],"prefix":"10.1371","volume":"21","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6295-8578","authenticated-orcid":true,"given":"Khanh Giang","family":"Le","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2026,4,28]]},"reference":[{"key":"pone.0347873.ref001","volume-title":"Global status report on road safety 2023","author":"World Health Organization","year":"2023"},{"key":"pone.0347873.ref002","doi-asserted-by":"crossref","first-page":"105949","DOI":"10.1016\/j.ssci.2022.105949","article-title":"Data-driven approaches for road safety: a comprehensive systematic literature review","volume":"158","author":"A Sohail","year":"2023","journal-title":"Saf Sci"},{"issue":"1","key":"pone.0347873.ref003","doi-asserted-by":"crossref","first-page":"19","DOI":"10.1007\/s41062-025-02409-9","article-title":"A state-of-the-art review of injury severity analysis in traffic crashes: Toward a generalized modeling framework","volume":"11","author":"M Awad","year":"2026","journal-title":"Innov Infrastruct Solut"},{"issue":"1","key":"pone.0347873.ref004","doi-asserted-by":"crossref","first-page":"107","DOI":"10.3390\/su16010107","article-title":"Urban traffic accident features investigation to improve urban transportation infrastructure sustainability by integrating GIS and data mining techniques","volume":"16","author":"KG Le","year":"2023","journal-title":"Sustainability"},{"issue":"13","key":"pone.0347873.ref005","doi-asserted-by":"crossref","first-page":"7115","DOI":"10.3390\/app15137115","article-title":"Road crash analysis and modeling: a systematic review of methods, data, and emerging technologies","volume":"15","author":"L Skaug","year":"2025","journal-title":"Appl Sci"},{"key":"pone.0347873.ref006","doi-asserted-by":"crossref","first-page":"237","DOI":"10.1016\/j.ijtst.2024.11.009","article-title":"Temporal dynamics of pedestrian injury severity: a seasonally constrained random parameters approach","volume":"20","author":"MA Abdulrazaq","year":"2024","journal-title":"Int J Transp Sci Technol"},{"key":"pone.0347873.ref007","doi-asserted-by":"crossref","first-page":"108277","DOI":"10.1016\/j.aap.2025.108277","article-title":"Seasonal instability in the determinants of vulnerable road user crashes: a partially temporally constrained modeling approach","volume":"224","author":"MA Abdulrazaq","year":"2026","journal-title":"Accid Anal Prev"},{"key":"pone.0347873.ref008","first-page":"1","article-title":"Unobserved heterogeneity and the statistical analysis of highway accident data","volume":"11","author":"FL Mannering","year":"2016","journal-title":"Anal Methods Accid Res"},{"key":"pone.0347873.ref009","doi-asserted-by":"crossref","first-page":"27","DOI":"10.1016\/j.aap.2017.08.008","article-title":"Comparison of four statistical and machine learning methods for crash severity prediction","volume":"108","author":"A Iranitalab","year":"2017","journal-title":"Accid Anal Prev"},{"key":"pone.0347873.ref010","doi-asserted-by":"crossref","first-page":"60079","DOI":"10.1109\/ACCESS.2018.2874979","article-title":"Comparing prediction performance for crash injury severity among various machine learning and statistical methods","volume":"6","author":"J Zhang","year":"2018","journal-title":"IEEE Access"},{"key":"pone.0347873.ref011","first-page":"129","article-title":"Improving road safety: supervised machine learning analysis of factors influencing crash severity","volume":"127","author":"KG Le","year":"2025","journal-title":"Sci J Silesian Univ Technol Ser Transp"},{"key":"pone.0347873.ref012","doi-asserted-by":"crossref","first-page":"106090","DOI":"10.1016\/j.aap.2021.106090","article-title":"A deep learning based traffic crash severity prediction framework","volume":"154","author":"MA Rahim","year":"2021","journal-title":"Accid Anal Prev"},{"issue":"4","key":"pone.0347873.ref013","doi-asserted-by":"crossref","first-page":"408","DOI":"10.1080\/17457300.2021.1928233","article-title":"Injury severity prediction of traffic crashes with ensemble machine learning techniques: a comparative study","volume":"28","author":"A Jamal","year":"2021","journal-title":"Int J Inj Contr Saf Promot"},{"key":"pone.0347873.ref014","doi-asserted-by":"crossref","first-page":"107271","DOI":"10.1016\/j.aap.2023.107271","article-title":"Crash injury severity prediction considering data imbalance: A Wasserstein generative adversarial network with gradient penalty approach","volume":"192","author":"Y Li","year":"2023","journal-title":"Accid Anal Prev"},{"issue":"7","key":"pone.0347873.ref015","doi-asserted-by":"crossref","first-page":"61","DOI":"10.3390\/infrastructures5070061","article-title":"Handling imbalanced data in road crash severity prediction by machine learning algorithms","volume":"5","author":"N Fiorentini","year":"2020","journal-title":"Infrastructures"},{"issue":"3","key":"pone.0347873.ref016","doi-asserted-by":"crossref","first-page":"707","DOI":"10.1007\/s13385-025-00428-5","article-title":"From point to probabilistic gradient boosting for claim frequency and severity prediction","volume":"15","author":"D Chevalier","year":"2025","journal-title":"Eur Actuar J"},{"key":"pone.0347873.ref017","unstructured":"Sinaga KP, Nair AS. Calibration meets reality: Making machine learning predictions trustworthy. arXiv [Preprint]. arXiv:2509.23665. 2025; doi: 10.48550\/arXiv.2509.23665"},{"issue":"6","key":"pone.0347873.ref018","doi-asserted-by":"crossref","first-page":"855","DOI":"10.1080\/01441647.2021.1954108","article-title":"Applications of machine learning methods in traffic crash severity modelling: Current status and future directions","volume":"41","author":"X Wen","year":"2021","journal-title":"Transp Rev"},{"key":"pone.0347873.ref019","doi-asserted-by":"crossref","first-page":"103969","DOI":"10.1016\/j.tre.2025.103969","article-title":"An interpretable machine learning framework for enhancing road transportation safety","volume":"195","author":"I Abdulrashid","year":"2025","journal-title":"Transp Res Part E Logist Transp Rev"},{"key":"pone.0347873.ref020","unstructured":"Department for Transport. Reported road casualties in Great Britain: STATS19 data. London: Department for Transport; 2025. https:\/\/www.gov.uk\/government\/collections\/road-accidents-and-safety-statistics"},{"issue":"7","key":"pone.0347873.ref021","first-page":"273","article-title":"Road safety in Great Britain: an exploratory data analysis","volume":"17","author":"JK Choudhary","year":"2023","journal-title":"Int J Transp Veh Eng"},{"key":"pone.0347873.ref022","volume-title":"Statistical analysis with missing data","author":"RJ Little","year":"2019","edition":"3"},{"issue":"4","key":"pone.0347873.ref023","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/2382577.2382579","article-title":"Leakage in data mining: formulation, detection, and avoidance","volume":"6","author":"S Kaufman","year":"2012","journal-title":"ACM Trans Knowl Discov Data"},{"key":"pone.0347873.ref024","doi-asserted-by":"crossref","DOI":"10.1007\/978-0-387-84858-7","volume-title":"The elements of statistical learning","author":"T Hastie","year":"2009","edition":"2"},{"key":"pone.0347873.ref025","first-page":"1137","volume-title":"A study of cross-validation and bootstrap for accuracy estimation and model selection","author":"R Kohavi","year":"1995"},{"issue":"5","key":"pone.0347873.ref026","doi-asserted-by":"crossref","first-page":"1666","DOI":"10.1016\/j.aap.2011.03.025","article-title":"The statistical analysis of highway crash-injury severities: a review and assessment of methodological alternatives","volume":"43","author":"PT Savolainen","year":"2011","journal-title":"Accid Anal Prev"},{"key":"pone.0347873.ref027","first-page":"3149","article-title":"LightGBM: a highly efficient gradient boosting decision tree","author":"G Ke","year":"2017","journal-title":"Adv Neural Inf Process Syst"},{"issue":"5","key":"pone.0347873.ref028","doi-asserted-by":"crossref","first-page":"1189","DOI":"10.1214\/aos\/1013203451","article-title":"Greedy function approximation: A gradient boosting machine","volume":"29","author":"JH Friedman","year":"2001","journal-title":"Ann Stat"},{"key":"pone.0347873.ref029","doi-asserted-by":"crossref","first-page":"207","DOI":"10.1016\/j.aap.2016.06.004","article-title":"Evaluation of the predictability of real-time crash risk models","volume":"94","author":"C Xu","year":"2016","journal-title":"Accid Anal Prev"},{"issue":"2","key":"pone.0347873.ref030","doi-asserted-by":"crossref","first-page":"230","DOI":"10.1518\/001872097778543886","article-title":"Humans and automation: use, misuse, disuse, abuse","volume":"39","author":"R Parasuraman","year":"1997","journal-title":"Hum Factors"},{"issue":"8","key":"pone.0347873.ref031","article-title":"Crash severity analysis of vulnerable road users using machine learning","volume":"16","author":"MMR Komol","year":"2021","journal-title":"PLoS One"},{"key":"pone.0347873.ref032","doi-asserted-by":"crossref","first-page":"1905","DOI":"10.1016\/j.procs.2024.09.704","article-title":"An explainable machine learning approach to traffic accident fatality prediction","volume":"246","author":"MdA Khan Rifat","year":"2024","journal-title":"Procedia Computer Science"},{"issue":"11","key":"pone.0347873.ref033","doi-asserted-by":"crossref","first-page":"3091","DOI":"10.1007\/s10439-020-02686-8","article-title":"An improved method for developing injury risk curves using the brier metric score","volume":"49","author":"ZS Hostetler","year":"2021","journal-title":"Ann Biomed Eng"},{"issue":"2","key":"pone.0347873.ref034","doi-asserted-by":"crossref","first-page":"185","DOI":"10.1002\/met.21","article-title":"Performance targets and the Brier score","volume":"14","author":"MS Roulston","year":"2007","journal-title":"Meteorol Appl"},{"issue":"4","key":"pone.0347873.ref035","doi-asserted-by":"crossref","first-page":"209","DOI":"10.1111\/j.1539-6924.1982.tb01384.x","article-title":"The theory of risk homeostasis: implications for safety and health","volume":"2","author":"GJS Wilde","year":"1982","journal-title":"Risk Anal"},{"key":"pone.0347873.ref036","first-page":"1","article-title":"Analytic methods in accident research: Methodological frontier and future directions","volume":"1","author":"FL Mannering","year":"2014","journal-title":"Anal Methods Accid Res"},{"issue":"6","key":"pone.0347873.ref037","doi-asserted-by":"crossref","first-page":"677","DOI":"10.1080\/01441640903414470","article-title":"Safety effects of winter weather: the state of knowledge and remaining challenges","volume":"30","author":"CK Strong","year":"2010","journal-title":"Transp Rev"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0347873","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,28]],"date-time":"2026-04-28T17:48:34Z","timestamp":1777398514000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0347873"}},"subtitle":[],"editor":[{"given":"Xuecai","family":"Xu","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,4,28]]},"references-count":37,"journal-issue":{"issue":"4","published-online":{"date-parts":[[2026,4,28]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0347873","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4,28]]}}}