A Study of Traffic Accidents in Spanish Intercity Roads by Means of Feature Vectors

A Study of Traffic Accidents in Spanish Intercity Roads by Means of Feature Vectors

D. Úbeda A. Gil L. Payá O. Reinoso 

Department of Systems Engineering and Automation, University Miguel Hernández de Elche, Spain

| |
| | Citation



Frequently, road traffic accidents are modelled as discrete and independent random and rare events, which possess a low probability of occurrence through time. Nevertheless, in order to study each accident individually it is necessary to obtain details of a number of characteristics that surround it, which may be correlated with each other. In this article, we propose to associate the probability of occurrence of an accident with a large number of features such as weather conditions, incidents caused by the start and end of a roadwork, geographical location of speed control radars, roadway infrastructure, etc. The influence of these features is significant and should be taken into account when proposing measures to help alleviate these undesirable events. The big data methods employed to extract the variables or features allow us to compose a series of vectors that will serve as a basis to study road accident distributions.


road traffic accident, road traffic data mining, weather features vectors


[1] European Union, Smart seat and seatbelt to help sleepy drivers stay alert. Research*eu Results Magazine, 42, pp. 6–7, 2015.

[2] Sabey, B.E. & Taylor, H., The known risks we run: the highway. Societal Risk Assessment, ed. R.C. Schwing,. & Albers, W.A., Springer US: Boston, pp. 43–70, 1980. http://dx.doi.org/10.1007/978-1-4899-0445-4_3

[3] Hakim, S., Shefer, D., Hakkert, A., & Hocherman, I., A critical review of macro models for road accidents. Accident Analysis and Prevention, ed. Abdel-Aty, M., Elsevier, pp. 379–400, 1991.

[4] Wang, C., Quddus, M.A. & Ison, S.G., Predicting accident frequency at their security levels and its application in site ranking using a two-stage mixed multivariable model. Accident Analysis and Prevention, ed. Abdel-Aty, M., Elsevier, pp. 1979–1990, 2011.

[5] Simoncic, M., A bayesian network model of two-car accidents. Journal of Transportation and Statistics, ed. Jeeves, A., United States Bureau of Transportation Statistics, pp. 13–25, 2004.

[6] Davison, P.A., Interrelationships between british drivers’ visual Abilities, age and road accidents histories. Ophthalmic and Physiological Optics, ed. Oxford Pergamon Press, pp. 195–204, 1995.

[7] Zegeer, C.V., Reinfurt, D., Hummer, J., Herf, L. & Hunter, W., Safety effects of cross-section design for two-lane roads. Transportation Research Board, ed. US Department of Transportation, 1987.

[8] Kalokota, K., Seneviratne, P.N., y Center, U.T., Accident Prediction Models for two-lane Rural Highways, Utah Transportation Center, 1994.

[9] Shankar, V., Milton, J. & Mannering, F., Modeling frequencies as zero-altered probability processes: an empirical inquiry. Accident Analysis and Prevention, ed. Abdel-Aty, M., Elsevier, pp. 829–837, 1997.

[10] DGT Statistical website, available at https://sedeapl.dgt.gob.es/WEB_IEST_CONSULTA/

[11] DGT Infocar, available at http://infocar.dgt.es/etraffic/

[12] Apache Cassandra, available at http://cassandra.apache.org/

[13] Wunderground, available at http://www.wunderground.com/