International Journal of Automotive Engineering
Online ISSN : 2185-0992
Print ISSN : 2185-0984
Research Paper
Factors Influencing Specificity and Sensitivity of Injury Severity Prediction (ISP) Algorithm for AACN
Chinmoy PalTomosaburo OkabeVimalathithan KulothunganNarahari SangollaJeyabharath ManoharanWang StewartJohn Combest
Author information
JOURNALS OPEN ACCESS

2016 Volume 7 Issue 1 Pages 15-22

Details
Abstract

To improve the accuracy of Injury Severity Prediction in the event of vehicle crash, a new methodology is proposed using the US vehicle accident database (NASS-CDS). This proposed method is an extension of the base algorithm introduced by Kononen et al. in which, some of the additional variables were introduced and branched logistic regression methodology was used. Results suggest that the proposed branching method has some advantage over the base algorithm due to better linearization of the complex multidimensional non-linear relationship of the input and output variables.

Information related to the author
© 2016 Society of Automotive Engineers of Japan, Inc
Previous article Next article
feedback
Top