Study of the evaluation of neural networks for the classification of the level of injury in traffic accidents (Original)
Keywords:
classification; artificial neural networks; unbalanced classes; performanceAbstract
In this article, the modeling of three neural networks, MLP, RBF and PNN is developed and studied. The objective of this work is the study of the evaluation of neural networks to classify the level of passenger injury of vehicles involved in traffic accidents. Several methods are used that will allow theoretical-practical deepening. In the research, the information about all people involved in car accidents on US roads during the year 2001 is taken as the field of action. The parameters used for neural networks are modeled. A comparative analysis of these networks was done in order to provide the best performance in the classification of the analyzed problem. It is concluded that the study guarantees to evaluate the performance of the networks used.
Downloads
References
Carvalho, C. B. (2000). Redes Neurais Artificiais: Teoria e Aplicações. Rio de Janeiro.
Chawla, N., Hall, L., Bowyer, K., & Kegelmeyer, W. (2002). SMOTE: Synthetic. Journal of Artificial Intelligence Research, 16, 321-357.
García, M., Villuendas, Y., Medina, M., Martínez, J., & Ruiz, J. (2008). Selección y construcción de objetos para el mejoramiento de un clasificador supervisado: un análisis crítico. Havana: Centav.
Haibo , H., Yang , B., Garcia, E., & Li, S. (2008). DASYN: Adaptive Synthetic Sampling Approach for Imbalanced. International Joint Conference on Neural Networks (IJCNN 2008), 1322-1329.
He, H., & Garcia, E. (2009). Learning from Imbalanced Data. IEEE.
JL Murray, C., D Lopez, A., Mathers, C., & Stein, C. (2001). The Global Burden of Disease 2000. World Health Organization. Fonte: http://www.who.int/healthinfo/paper36.pdf
Kotsiantis, S., & Pintelas, P. (2003). Mixture of Expert Agents for Handling Imbalanced Data Sets. Annals of Mathematics, Computing and TeleInformatics, 1, 46-55.
Kubat, M., & Matwin, S. (1997). Addressing the curse of imbalanced training sets: One sided selection. In Proceedings of the Fourteenth International Conference on Machine Learning, 179-186.
Peden, M. (2004). World report on road traffic injury prevention. World Health Organization. Fonte: http://www.who.int/healthinfo/paper36.pdf
Spinelli, A. (2010). Um estudo comparativo de métodos para balanceamento do conjunto de treinamento em aprendizado de redes neurais artificiais.
Vellasco, M. (2018). Presentações das aulas de Redes Neurais.
World Health , O. (2008). World health statistics 2008. World Health Organization. Geneve. Fonte: http://www.who.int/gho/publications/world_health_statistics/EN_WHS08_Full.pdf
World Health, O. (2008). The Global Burden of Disease: 2004 update. Organization, World Health. Geneve. Fonte: http://www.who.int/healthinfo/global_burden_disease/GBD_report_2004update_full.pdf?ua=1
