| With the development of economy in China freeway mileages is increasing rapidly. According to statistics, By the end of 2015, our country’s freeway mileage has come to over 120000 kilometers and has been ranked first in the world. The freeway has the advantages of high speed, large capacity and low transportation cost, which is important for the development of the economy along the line.However, because of the fully enclosed, interchange, to traffic separation and other special traffic environment, freeway accidents of rear end collision risk is much higher than that of the other road, because of high speed and large traffic flow, the freeway once happen traffic accident, the consequences are often very serious. It is significant to research on the influence factors of accident severity and put forward the countermeasures.In this context, this paper through an analysis of 407 freeway rear-end accidents, establish the severity of the freeway rear-end collison prediction model based on the cumulative Logistic regression models and the severity of the freeway rear-end collisson prediction model based on the support vector machine. Based on the accumulative Logistic model, we use mixed stepwise selection method and select the driving experience, the condition of vehicle safety, terrain, road conditions, weather, visibility were associated with a significant rear-end accident severity. Based on the SVM model, we select the driving experience, the condition of vehicle safety, overload, road conditions, road alignment, weather, visibility were associated with a significant rear-end accident severity. Finally, this paper based on the cumulative logistic regression and SVM model analysis results, from the driver, vehicle, road and environment to punish the influence degree, and puts forward the corresponding countermeasures to prevent the occurrence of rear end collision accidents on Expressway and reduce the severity of the crash... |
PDF Full Text Request