Comparing Explainability Frameworks for Accident Severity Prediction
Resumo
A predição da gravidade de sinistros com bases administrativas é limitada pela ausência de variáveis comportamentais, tornando mais relevante identificar quais variáveis são consistentemente associadas à gravidade entre métodos de explicabilidade do que maximizar a acurácia. Este artigo compara SHapley Additive exPlanations (SHAP) e Efeitos Marginais Médios (AMEs) em diferentes classificadores na base de sinistros em rodovias federais brasileiras (2017–2025). Os resultados revelam um núcleo estável de variáveis, concordância mais forte entre combinações baseadas em SHAP, e concordância mais heterogênea, de baixa a moderada, entre SHAP e AMEs.Referências
Aarts, L. and Van Schagen, I. (2006). Driving speed and the risk of road crashes: A review. Accident Analysis & Prevention, 38(2):215–224.
Anselin, L. (1988). Spatial econometrics: methods and models, volume 4. Springer Science & Business Media.
Brasil (1997). Código de trânsito brasileiro (lei nº 9.503, de 23 de setembro de 1997). Disponível em: [link]. Acesso em: 2 out. 2024.
Breiman, L. (2001). Random forests. Machine learning, 45(1):5–32.
Bucsuházy, K., Matuchová, E., Zvala, R., Moravcová, P., Kostíková, M., and Mikulec, R. (2020). Human factors contributing to the road traffic accident occurrence. Transportation research procedia, 45:555–561.
Cameron, A. (2005). Microeconometrics: methods and applications. Cambridge University.
Carvalho, C. H. R. d. (2020). Custos dos acidentes de trânsito no brasil: Estimativa simplificada com base na atualização das pesquisas do ipea sobre custos de acidentes nos aglomerados urbanos e rodovias.
Chen, J., Liu, P., Wang, S., Zheng, N., and Guo, X. (2025). Prediction and interpretation of crash severity using machine learning based on imbalanced traffic crash data. Journal of Safety Research, 93:185–199.
Cox, D. R. (1958). The regression analysis of binary sequences. Journal of the Royal Statistical Society Series B: Statistical Methodology, 20(2):215–232.
Departamento Nacional de Infraestrutura de Transportes (DNIT) (2024). Controle de velocidade — dados abertos dnit. [link]. Acesso em: 02 out. 2024.
Dong, S., Khattak, A., Ullah, I., Zhou, J., and Hussain, A. (2022). Predicting and analyzing road traffic injury severity using boosting-based ensemble learning models with shapley additive explanations. International journal of environmental research and public health, 19(5):2925.
Dormann, C. F., Elith, J., Bacher, S., Buchmann, C., Carl, G., Carré, G., Marquéz, J. R. G., Gruber, B., Lafourcade, B., Leitão, P. J., et al. (2013). Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. Ecography, 36(1):27–46.
Efron, B. and Tibshirani, R. J. (1994). An introduction to the bootstrap. Chapman and Hall/CRC.
Franceschi, L., Kaesemodel, L., de Vargas, V. d. C. C., Konrath, A. C., Nakamura, L. R., Ramires, T. G., Barreto, C. B. M., and Valente, A. M. (2022). Factors related to highway crash severity in brazil through a multinomial logistic regression model. Transportes, 30(1):2566–2566.
Hameed, I. and Roth, A. (2024). Statistical significance of feature importance rankings. arXiv preprint arXiv:2401.15800.
Harrell, F. E. et al. (2001). Regression modeling strategies: with applications to linear models, logistic regression, and survival analysis, volume 608. Springer.
Hastie, T., Tibshirani, R., Friedman, J. H., and Friedman, J. H. (2009). The elements of statistical learning: data mining, inference, and prediction, volume 2. Springer.
Kendall, M. G. (1938). A new measure of rank correlation. Biometrika, 30(1-2):81–93.
Khanum, H., Garg, A., Faheem, M. I., and Kulkarni, R. (2025). A methodological framework for road accident severity prediction for indian highways using machine learning models. MethodsX, page 103728.
Lundberg, S. M., Erion, G., Chen, H., DeGrave, A., Prutkin, J. M., Nair, B., Katz, R., Himmelfarb, J., Bansal, N., and Lee, S.-I. (2020). From local explanations to global understanding with explainable ai for trees. Nature machine intelligence, 2(1):56–67.
Lundberg, S. M. and Lee, S.-I. (2017). A unified approach to interpreting model predictions. Advances in neural information processing systems, 30.
Mafort, E. V. and Kappel, M. A. (2025). Técnicas de aprendizado de máquina para predição de gravidade de acidentes em rodovias do estado do rio de janeiro. Revista Eletrônica de Iniciação Científica em Computação, 23:244–252.
Mannering, F., Bhat, C. R., Shankar, V., and Abdel-Aty, M. (2020). Big data, traditional data and the tradeoffs between prediction and causality in highway-safety analysis. Analytic methods in accident research, 25:100113.
Micci-Barreca, D. (2001). A preprocessing scheme for high-cardinality categorical attributes in classification and prediction problems. ACM SIGKDD explorations newsletter, 3(1):27–32.
Molnar, C., König, G., Herbinger, J., Freiesleben, T., Dandl, S., Scholbeck, C. A., Casalicchio, G., Grosse-Wentrup, M., and Bischl, B. (2020). Pitfalls to avoid when interpreting machine learning models.
Polícia Rodoviária Federal (2026). Dados abertos da prf. [link]. Acesso em: jan. 2026.
Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., and Gulin, A. (2018). Cat-boost: unbiased boosting with categorical features. Advances in neural information processing systems, 31.
Saarela, M. and Jauhiainen, S. (2021). Comparison of feature importance measures as explanations for classification models. SN Applied Sciences, (2).
SHAP developers (2026). shap.treeexplainer — SHAP latest documentation. [link]. Accessed: 2026-08-14.
Shapley, L. S. et al. (1953). A value for n-person games.
Shmueli, G. (2010). To explain or to predict? Statistical science, pages 289–310.
Spearman, C. (1987). The proof and measurement of association between two things. The American journal of psychology, 100(3/4):441–471.
Swamy, V., Radmehr, B., Krco, N., Marras, M., and Käser, T. (2022). Evaluating the explainers: black-box explainable machine learning for student success prediction in moocs. arXiv preprint arXiv:2207.00551.
Wang, C. and Serre, T. (2025). A hybrid approach to investigating factors associated with crash injury severity: Integrating interpretable machine learning with logit model. Applied Sciences, 15(19):10417.
Webber, W., Moffat, A., and Zobel, J. (2010). A similarity measure for indefinite rankings. ACM Transactions on Information Systems (TOIS), 28(4):1–38.
Wen, X., Xie, Y., Jiang, L., Pu, Z., and Ge, T. (2021). Applications of machine learning methods in traffic crash severity modelling: current status and future directions. Transport reviews, 41(6):855–879.
Wilson, C., Willis, C., Hendrikz, J. K., Le Brocque, R., and Bellamy, N. (2010). Speed cameras for the prevention of road traffic injuries and deaths. Cochrane database of systematic reviews, (11).
Zhang, X., Wen, H., Yamamoto, T., and Zeng, Q. (2021). Investigating hazardous factors affecting freeway crash injury severity incorporating real-time weather data: Using a bayesian multinomial logit model with conditional autoregressive priors. Journal of safety research, 76:248–255.
Anselin, L. (1988). Spatial econometrics: methods and models, volume 4. Springer Science & Business Media.
Brasil (1997). Código de trânsito brasileiro (lei nº 9.503, de 23 de setembro de 1997). Disponível em: [link]. Acesso em: 2 out. 2024.
Breiman, L. (2001). Random forests. Machine learning, 45(1):5–32.
Bucsuházy, K., Matuchová, E., Zvala, R., Moravcová, P., Kostíková, M., and Mikulec, R. (2020). Human factors contributing to the road traffic accident occurrence. Transportation research procedia, 45:555–561.
Cameron, A. (2005). Microeconometrics: methods and applications. Cambridge University.
Carvalho, C. H. R. d. (2020). Custos dos acidentes de trânsito no brasil: Estimativa simplificada com base na atualização das pesquisas do ipea sobre custos de acidentes nos aglomerados urbanos e rodovias.
Chen, J., Liu, P., Wang, S., Zheng, N., and Guo, X. (2025). Prediction and interpretation of crash severity using machine learning based on imbalanced traffic crash data. Journal of Safety Research, 93:185–199.
Cox, D. R. (1958). The regression analysis of binary sequences. Journal of the Royal Statistical Society Series B: Statistical Methodology, 20(2):215–232.
Departamento Nacional de Infraestrutura de Transportes (DNIT) (2024). Controle de velocidade — dados abertos dnit. [link]. Acesso em: 02 out. 2024.
Dong, S., Khattak, A., Ullah, I., Zhou, J., and Hussain, A. (2022). Predicting and analyzing road traffic injury severity using boosting-based ensemble learning models with shapley additive explanations. International journal of environmental research and public health, 19(5):2925.
Dormann, C. F., Elith, J., Bacher, S., Buchmann, C., Carl, G., Carré, G., Marquéz, J. R. G., Gruber, B., Lafourcade, B., Leitão, P. J., et al. (2013). Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. Ecography, 36(1):27–46.
Efron, B. and Tibshirani, R. J. (1994). An introduction to the bootstrap. Chapman and Hall/CRC.
Franceschi, L., Kaesemodel, L., de Vargas, V. d. C. C., Konrath, A. C., Nakamura, L. R., Ramires, T. G., Barreto, C. B. M., and Valente, A. M. (2022). Factors related to highway crash severity in brazil through a multinomial logistic regression model. Transportes, 30(1):2566–2566.
Hameed, I. and Roth, A. (2024). Statistical significance of feature importance rankings. arXiv preprint arXiv:2401.15800.
Harrell, F. E. et al. (2001). Regression modeling strategies: with applications to linear models, logistic regression, and survival analysis, volume 608. Springer.
Hastie, T., Tibshirani, R., Friedman, J. H., and Friedman, J. H. (2009). The elements of statistical learning: data mining, inference, and prediction, volume 2. Springer.
Kendall, M. G. (1938). A new measure of rank correlation. Biometrika, 30(1-2):81–93.
Khanum, H., Garg, A., Faheem, M. I., and Kulkarni, R. (2025). A methodological framework for road accident severity prediction for indian highways using machine learning models. MethodsX, page 103728.
Lundberg, S. M., Erion, G., Chen, H., DeGrave, A., Prutkin, J. M., Nair, B., Katz, R., Himmelfarb, J., Bansal, N., and Lee, S.-I. (2020). From local explanations to global understanding with explainable ai for trees. Nature machine intelligence, 2(1):56–67.
Lundberg, S. M. and Lee, S.-I. (2017). A unified approach to interpreting model predictions. Advances in neural information processing systems, 30.
Mafort, E. V. and Kappel, M. A. (2025). Técnicas de aprendizado de máquina para predição de gravidade de acidentes em rodovias do estado do rio de janeiro. Revista Eletrônica de Iniciação Científica em Computação, 23:244–252.
Mannering, F., Bhat, C. R., Shankar, V., and Abdel-Aty, M. (2020). Big data, traditional data and the tradeoffs between prediction and causality in highway-safety analysis. Analytic methods in accident research, 25:100113.
Micci-Barreca, D. (2001). A preprocessing scheme for high-cardinality categorical attributes in classification and prediction problems. ACM SIGKDD explorations newsletter, 3(1):27–32.
Molnar, C., König, G., Herbinger, J., Freiesleben, T., Dandl, S., Scholbeck, C. A., Casalicchio, G., Grosse-Wentrup, M., and Bischl, B. (2020). Pitfalls to avoid when interpreting machine learning models.
Polícia Rodoviária Federal (2026). Dados abertos da prf. [link]. Acesso em: jan. 2026.
Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., and Gulin, A. (2018). Cat-boost: unbiased boosting with categorical features. Advances in neural information processing systems, 31.
Saarela, M. and Jauhiainen, S. (2021). Comparison of feature importance measures as explanations for classification models. SN Applied Sciences, (2).
SHAP developers (2026). shap.treeexplainer — SHAP latest documentation. [link]. Accessed: 2026-08-14.
Shapley, L. S. et al. (1953). A value for n-person games.
Shmueli, G. (2010). To explain or to predict? Statistical science, pages 289–310.
Spearman, C. (1987). The proof and measurement of association between two things. The American journal of psychology, 100(3/4):441–471.
Swamy, V., Radmehr, B., Krco, N., Marras, M., and Käser, T. (2022). Evaluating the explainers: black-box explainable machine learning for student success prediction in moocs. arXiv preprint arXiv:2207.00551.
Wang, C. and Serre, T. (2025). A hybrid approach to investigating factors associated with crash injury severity: Integrating interpretable machine learning with logit model. Applied Sciences, 15(19):10417.
Webber, W., Moffat, A., and Zobel, J. (2010). A similarity measure for indefinite rankings. ACM Transactions on Information Systems (TOIS), 28(4):1–38.
Wen, X., Xie, Y., Jiang, L., Pu, Z., and Ge, T. (2021). Applications of machine learning methods in traffic crash severity modelling: current status and future directions. Transport reviews, 41(6):855–879.
Wilson, C., Willis, C., Hendrikz, J. K., Le Brocque, R., and Bellamy, N. (2010). Speed cameras for the prevention of road traffic injuries and deaths. Cochrane database of systematic reviews, (11).
Zhang, X., Wen, H., Yamamoto, T., and Zeng, Q. (2021). Investigating hazardous factors affecting freeway crash injury severity incorporating real-time weather data: Using a bayesian multinomial logit model with conditional autoregressive priors. Journal of safety research, 76:248–255.
Publicado
19/10/2026
Como Citar
RODRIGUES, Diego Souza; LOPES, Rodolfo Ayala; SILVA, Rodrigo.
Comparing Explainability Frameworks for Accident Severity Prediction. In: ENCONTRO NACIONAL DE INTELIGÊNCIA ARTIFICIAL E COMPUTACIONAL (ENIAC), 23. , 2026, Cuiabá/MT.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2026
.
p. 1004-1015.
ISSN 2763-9061.
DOI: https://doi.org/10.5753/eniac.2026.30778.
