USF St. Petersburg campus Faculty Publications
Classifying injuries occurrence in motor vehicle collisions using artificial neural network.
Document Type
Presentation
Publication Date
2011
ISSN
1541-9312
Abstract
Vehicle collisions amount to a significant loss of life in America. This study used artificial neural networks as a means to predict the occurrence of injury of a vehicle collision. Using Neural Ware’s Predict software a neural network structure was trained, tested, and validated using data from the 2006 and 2007 Florida Traffic Crash Database. The objective was to assess whether or not properly designed neural network architecture could adequately classify the levels of the “Injury Occurrence” output variable, given certain inputs such as demographic and environmental factors involved in crashes. A Kolmogorov-Smirnov statistical analysis was employed to objectively assess whether or not the neural network properly classified the levels of Injury Occurrence and to what extent. The artificial network’s computational power was iteratively increased by adding hidden layers thus boosting its performance. A sensitivity analysis was used to find the level of contribution the input variables had on the “Injury Occurrence” output variable. Top three positive and negative most impacting factors were identified and the implications were discussed at the end of the paper.
Language
en_US
Publisher
Sage
Recommended Citation
Liu, D., Soloman, D., & Hardy, L. (2011). Classifying injuries occurrence in motor vehicle collisions using artificial neural network. In Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 55(1), 1808-1812. doi: 10.1177/1071181311551376.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Comments
Abstract only. Full-text article is available through licensed access provided by the publisher. Published in Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 55(1), 1808-1812. doi: 10.1177/1071181311551376. Members of the USF System may access the full-text of the article through the authenticated link provided.