access icon free Automatic classification of traffic incident's severity using machine learning approaches

During daily work at a Transport Management Centre (TMC), the operators have to record and process a large volume of traffic information especially incident records. Their tasks involve manual classification of the data and then decide appropriate operations to clear the incidents on time. A real-time automatic decision support system can minimise an operator's responded time and hence reduce congestion. Besides standard descriptions (e.g. incident location, date, time, lanes affected), severity is an important criteria that operators have to evaluate based on all available information before any control commands can be issued. The NSW TMC and the research organisation Data61 in Sydney have collaborated to discover and visualise frequent patterns in historical incident response records, leading to the automatic classification of severity levels among past incidents using advanced machine learning, active learning and outlier detection techniques. The experiments were executed using 4 years TMC's incident logs from 2011 to 2014 which includes >40,000 records. The classification model achieved nearly 90% accuracy in five-fold cross-validation and is expected to help the TMC to improve its procedures, response plans, and resource allocations.

Inspec keywords: pattern classification; learning (artificial intelligence); traffic engineering computing

Other keywords: active learning; Transport Management Centre; machine learning; NSW TMC; historical incident response records; outlier detection techniques; traffic incident severity classification

Subjects: Traffic engineering computing; Knowledge engineering techniques; Data handling techniques

References

    1. 1)
      • 1. Taib, R., Yee, D., Fang, F., et al: ‘Improved incident management through anomaly detection in historical records’. Proc. of the 21th ITS World Congress, Detroit, USA, 2014, pp. 1123.
    2. 2)
      • 9. Rish, I.: ‘An empirical study of the naive Bayes classifier’. IJCAI 2001 workshop on empirical methods in artificial intelligence, IBM New York, 2001, vol. 3, no. 22, pp. 4146.
    3. 3)
      • 2. Nowakowska, M.: ‘Logistic models in crash severity classification based on road characteristics’. Transportation Research Record: Journal of the Transportation Research Board, No. 2148, Transportation Research Board of the National Academies, Washington, DC, 2010, pp. 1626.
    4. 4)
      • 17. Tong, S., Koller, D.: ‘Support vector machine active learning with applications to text classification’, J. Mach. Learn. Res., 2001, 2, pp. 4566.
    5. 5)
      • 4. Sohn, S.Y., Lee, S.H.: ‘Data fusion, ensemble and clustering to improve the classification accuracy for the severity of road traffic accidents in Korea’, Safety Science, 2003, 41, (1), pp. 114.
    6. 6)
      • 14. Hodge, V.J., Austin, J.: ‘A survey of outlier detection methodologies’, Artif. Intell. Rev., 2004, 22, (2), pp. 85126.
    7. 7)
      • 6. Miranda-Moreno, L., Fu, L., Ukkusuri, S., et al: ‘How to incorporate accident severity and vehicle occupancy into the hot spot identification process?’. Transport. Res. Record, Washington, D.C., 2009, pp. 5360.
    8. 8)
      • 10. Lewis, D.: ‘Naive (Bayes) at forty: the independence assumption in information retrieval’. Machine learning: ECML, Springer Berlin Heidelberg, 1998, pp. 415.
    9. 9)
      • 19. Roy, N., McCallum, A.: ‘Toward optimal active learning through Monte Carlo estimation of error reduction’. ICML, Williamstown, 2001, pp. 441448.
    10. 10)
      • 8. Hall, M., Frank, E., Holmes, G., et al: ‘The WEKA data mining software: an update’, ACM SIGKDD Explorations Newsletter, 2009, 11, (1), pp. 1018.
    11. 11)
      • 3. Sohn, S.Y., Shin, H.: ‘Pattern recognition for road traffic accident severity in Korea’, Ergonomics, 2001, 44, (1), pp. 107117.
    12. 12)
      • 18. Campbell, C., Cristianini, N., Smola, A.: ‘Query learning with large margin classifiers’. ICML, 2000, pp. 111118.
    13. 13)
      • 7. Chong, M., Abraham, A., Paprzycki, M.: ‘Traffic accident analysis using machine learning paradigms’, Informatica, 2005, 29, (1), pp. 8998.
    14. 14)
      • 11. Joachims, T.: ‘Text categorization with support vector machines: learning with many relevant features’. In Claire, N., Céline, R. (Eds.): (Springer, Berlin Heidelberg, 1998), pp. 137142.
    15. 15)
      • 21. Osugi, T., Kim, D., Scott, S.: ‘Balancing exploration and exploitation: a new algorithm for active machine learning’. Fifth IEEE Int. Conf. on Data Mining, IEEE, November 2005, p. 8–pp.
    16. 16)
      • 12. Quinlan, J.R.: ‘C4. 5: programs for machine learning’ (Elsevier, 2014).
    17. 17)
      • 5. Tesema, T.B., Abraham, A., Grosan, C.: ‘Rule mining and classification of road traffic accidents using adaptive regression trees’, Int. J. Simul., 2005, 6, (10), pp. 8094.
    18. 18)
      • 15. Settles, B.: ‘Active learning literature survey’. University of Wisconsin, Madison, 2010, 52, (55–66), 11.
    19. 19)
      • 16. Schohn, G., Cohn, D.: ‘Less is more: active learning with support vector machines’. ICML, 2000, pp. 839846.
    20. 20)
      • 20. Nguyen, D.H., Patrick, J.D.: ‘Supervised machine learning and active learning in classification of radiology reports’, J. Am. Med. Informatics Assoc., 2014, 21, (5), pp. 893901.
    21. 21)
      • 13. Kohavi, R.: ‘A study of cross-validation and bootstrap for accuracy estimation and model selection’. IJCAI, 1995, vol. 14, no. 2, pp. 11371145.
http://iet.metastore.ingenta.com/content/journals/10.1049/iet-its.2017.0051
Loading

Related content

content/journals/10.1049/iet-its.2017.0051
pub_keyword,iet_inspecKeyword,pub_concept
6
6
Loading