The hit-and-run (H-N-R) crashes are increasing in Vadodara. The purpose of this research is to investigate the pattern of H-N-R crashes and to determine the likelihood of a crash being a H-N-R crash. For the study, data was collected from the police department in the form of an FIR (first investigation report) and then extracted into a spreadsheet. To get insights about the data, exploratory analysis was performed, followed by χ2-test to find out the correlation between independent variables. Once the valid variables were identified, a binomial logistic regression model was developed to predict the probability of a crash being a H-N-R crash. From the model developed the influencing factors, confusion matrix, and ROC curve were obtained. It was found that the victim vehicle types, like four-wheelers, two-wheelers, and auto rickshaws, are at higher risk in cases of H-N-R crashes. The maneuvers of an impacting vehicle, like turning or reversing, were found to reduce the probability of a H-N-R crash with respect to going straight. It was also found that the victim’s gender and the number of victims also affected the H-N-R crash.
Analysis and Prediction of Hit-and-Run Road Accidents
Lecture Notes in Civil Engineering
International Conference on Transportation Infrastructure Projects : Conception to Execution ; 2022 ; Haridwar, India September 14, 2022 - September 17, 2022
2023-07-22
14 pages
Article/Chapter (Book)
Electronic Resource
English
TIBKAT | 2000
|Engineering Index Backfile | 1945
|Engineering Index Backfile | 1945
|Engineering Index Backfile | 1943
|Engineering Index Backfile | 1939
|