Analysis and Prediction of Passenger Injury Severity under the Influence of Clamping Based on Double-Layer Evaluation Model
Li Xuan, Han Tianyuan, Lü Kaiguang Qiao Jie, Liu Yongtao
Chang’an University, Xi’an 710064) (
Abstract In order to determine the factors influencing the severity of passenger injury under the influence of【 】clamping in a car crash, 714 sets of car crash data including 24 vehicle types are selected from Vehicle Physical Evidence Forensic Center Database of Chang’an University database. Through the Apriori association rules, the relevant characteristic factors are excavated, and the two-layer evaluation model is established to analyze the characteristic factors at the vehicle level and the passenger level. The results show that the tissue structure variable, the body part variable and the injury type variable are affected by the quality variable of the damaged component up to 28.89% , 15.79% and 14.03% respectively. Compared with other parts, the head injury is more serious in crash. The more serious injury is often accompanied by severe fractures and large- scale injuries of occupants. Finally, the random forest model is used to predict the severity of occupant injury, and the accuracy of the prediction results can reach 82.97%.
Key words: Clamping effect, Severity of injury, Apriori association rule, Two- layer evaluation model, Random forest, Analysis and prediction